Economics of Education Review 80 (2021) 102065

Contents lists available at ScienceDirect

Economics of Education Review

journal homepage: www.elsevier.com/locate/econedurev

No experience, no employment: The effect of vocational education and

training work experience on labour market outcomes after higher education
Maria Esther Oswald-Egg *, 1, Ursula Renold 1
Chair of Education Systems, D-MTEC, ETH Zurich, Leonhardstrasse 21, Zurich 8092, Switzerland

A R T I C L E I N F O A B S T R A C T

Keywords: Higher education graduates with work experience enter the labour market more smoothly. This study analyses
Work experience how work experience from vocational education and training (VET) affects labour market outcomes after higher
Higher education education. To account for selection into VET, we use the regional enrolment rate as an instrument for upper-
Vocational education and training
secondary VET. Results suggest that work experience gained during VET leads to significantly higher wages of
Labour market outcomes
Human capital
7% to 19% one year after graduation from higher education and two months less search time for first employment
Signalling but does not significantly lower the probability of an internship in the post-graduation year. However, these
Screening positive effects do not persist: the effect is no longer robustly significant for wages, unemployment or employment
Social network position after five years. The effect operates through the human capital (specific and general), screening and
signalling channels, not the social network channel. Our results suggest that upper-secondary VET is an equivalent
pathway to academic education, not merely the second-best, for individuals planning on higher education.

1. Introduction
Workers with experience are in high demand in the labour market
(Salvisberg, 2010) due to high productivity (Arrow, 1962) and low
training costs (Thurow, 1975). This demand for experienced workers is
the reason that institutions of higher education – such as universities and
universities of applied sciences (hereafter, ‘UAS’)2 – have begun
providing work experience in their curricula, including (mandatory)
internships (Cranmer, 2006; Helyer & Lee, 2014; Billett, 2014; Silva
et al., 2018). This work experience equips students with marketable
skills before they graduate and enter the labour market.
However, a number of empirical studies that account for self-
selection into work experience opportunities before graduation from
higher education find mixed results for labour market outcomes after
graduation. These studies typically analyse the effects of having an
internship, a student job, or both during higher education (San, 1986;
Ehrenberg & Sherman, 1987; Häkkinen, 2006; Klein & Weiss, 2011;
Weiss, Klein, & Grauenhorst, 2014; Baert, Rotsaert, Verhaest, & Omey,
2016; Nunley, Pugh, Romero, & Seals, 2016; Baert, Neyt, Omey, &
Verhaest, 2017; Margaryan, Saniter, Schumann, & Siedler, 2019; Bolli,
Caves, & Oswald-Egg, 2019a). Moreover, evidence on the effect of in­
ternships shows either no effect (Klein & Weiss, 2011; Weiss et al., 2014)
or a significantly positive one (Nunley et al., 2016; Margaryan et al.,
2019; Bolli et al., 2019a). Evidence on student jobs also shows mixed
results: some studies find no effect (Ehrenberg & Sherman, 1987;
Häkkinen, 2006; Baert et al., 2016), whereas others find positive effects
(San, 1986; Weiss et al., 2014). Two major drawbacks of gaining work
experience while studying – the time it takes away from studying and its
possible lack of relevance to the student’s field (Stern, Mc Million,
Hopkins, & Stone, 1990) – may account for these mixed results.
Although students who gain work experience before studying can
avoid these drawbacks, they suffer from the drawback of delayed
graduation. However, the evidence of the effect of pre-study work
experience on labour market outcomes after higher education is scarce.

* Corresponding author.
E-mail address: m.e.oswald-egg@mtec.ethz.ch (M.E. Oswald-Egg).
1
The authors thank Prof. Dr. Jan-Egbert Sturm, two anonymous reviewers, the members of the Swiss Leading House on the Economics of Education, and the
participants of the 2017 PhD Discussion Workshop on ’Personnel Economics and Economics of Education’, and the KOF Brown Bag Seminar for helpful comments
and suggestions. A special thanks goes to Natalie Reid and Katherine Caves Gahr for language consulting. All responsibility for remaining mistakes lies with the
authors.
2
UASs are institutions of higher education that teach and conduct applied research Lehnert and Pfister (2018). Pfister, Rinawi, Harhoff, and Backes-Gellner (2018,
p. 1) state that the name ‘Universities of Applied Sciences’ is used for Swiss institutions called ‘Fachhochschulen’ or for German institutions called ‘Fachhochschulen’
and more recently ‘Hochschulen für angewandte Wissenschaften (HAW)’.

https://doi.org/10.1016/j.econedurev.2020.102065
Received 6 January 2020; Received in revised form 26 September 2020; Accepted 1 November 2020
Available online 17 January 2021
0272-7757/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
M.E. Oswald-Egg and U. Renold
Weiss et al. (2014) analyse work experience from employment prior to
studying, finding no significant effects on wages, search time or class
positions. An as yet under-explored way of gaining work experience
before studying without delaying graduation is from vocational educa­
tion and training (VET), an upper-secondary education pathway lasting
three to four years and combining formal education at school with
curricula-driven practical training (OECD, 2004). In school-based VET,
the practical training takes place in classroom workshops, and in dual
VET the training takes place at a company, usually in the form of ap­
prenticeships (Wolter & Ryan, 2011). The likely advantage of VET,
especially dual VET, is that its graduates enter the labour market with
work experience and have the option of continuing to higher education.
The evidence on the impact of VET work experience on labour
market outcomes is also mixed. Shaw (2012) finds that upper-secondary
VET diplomas improve labour market outcomes after higher education
in the UK, while Agarwal, Brunello, and Rocco (2019) find that VET
work experience reduces search time after graduation but negatively
impacts wages, working hours and job positions in Italy. However,
neither the UK nor Italy has a great deal of dual VET, limiting the
amount of work experience that students can gain from those pro­
grammes. Similarly, their education systems are either not very
permeable,3 meaning that relatively few students progress from VET to
higher education (Pauline & Simon, 2013), or appear unlinked to the
needs of the labour market (OECD, 2017).
By focusing on VET work experience and its effect on labour market
outcomes after higher education, this study analyses the value of work
experience for higher education graduates. We focus on dual VET in a
permeable education system where students commonly progress from VET
to higher education. We use a case study in which dual VET predominates
over school-based VET, which has weaker labour market outcomes (Bolli,
Oswald-Egg, & Rageth, 2019b). This distinction is important in times when
policymakers are promoting VET, because neglecting the difference is likely
to lead to the erroneous conclusion that all types of VET constitute a
second-best option. Such a conclusion is especially harmful because dual
VET enhances students’ motivation to stay in education (Wolter & Ryan,
2011) and is not prone to the drawbacks of delayed graduation or
field-unrelated work experience. Moreover, little is known about the
channels through which the effects of VET work experience on labour
market outcomes after higher education flow. Yet knowing these channels is
critical for finding the best strategy for improving the youth labour market.
To analyse the effect of VET work experience on labour market out­
comes after higher education, we use Switzerland as a case study. Three
features of the Swiss labour market and education system make
Switzerland ideal for this analysis. First, the Swiss labour market functions
well and has for years had very low unemployment rates,4 so that highly
educated individuals can easily find work. Second, in the Swiss education
system VET at the upper-secondary level accommodates about two-thirds
of a cohort (SERI, 2020), roughly 90% of whom follow a dual VET pro­
gramme (SERI, 2020), which contains a substantial amount of workplace
training. Third, the Swiss education system is permeable, with open access
to two types of academic higher education institutions (universities and
UASs) after graduation from upper-secondary education.5
3
The definition of permeability from the European Centre for the Develop­
ment of Vocational Training is as follows: ‘Capacity of education and training
systems to enable learners to (1) access and move among different pathways
(programmes, levels) and systems; (2) validate learning outcomes acquired in
another system or in non-formal/informal settings’.
4
From 2010–2017 the unemployment rate (according to the ILO definition)
ranged from 3.5% to 5.5% (source: Federal Statistical Office of Switzerland –
Swiss Labour Force Survey, 4 May 2020).
5
Some additional requirements may apply, e.g. an additional day at voca­
tional baccalaureate school (parallel to the VET) or a university aptitude test for
students with a VET diploma, and a working year in the desired field of edu­
cation for students with an academic diploma.
2
Economics of Education Review 80 (2021) 102065
We exploit data from the Swiss graduate survey conducted in 2011,
2013, 2015 and 2017 on higher education graduates (HE graduates) one
year after graduation and in 2015 and 2017 on HE graduates five years
after graduation. These surveys provide data on their labour market
outcomes, such as wage, search time, internship, unemployment and
employment position. Moreover, these data allow us access to a rich
collection of variables about personal characteristics, education,
continuous education, work experience and employer characteristics.
Given the self-selection of HE graduates into obtaining an academic
diploma6 or a VET diploma at the upper-secondary level, we use two
estimation methods. In the first method, we include a large set of control
variables on observable characteristics. In the second method, to capture
unobservable differences, we use instrumental variable regressions. Our
instrument for having VET work experience is the regional7 enrolment
rate into VET. Specifically, we use the enrolment rate into VET when HE
graduates were 14, in the region where they were living before starting
higher education. Because HE graduates are mobile after graduation
(Oggenfuss & Wolter, 2018), the regional enrolment rate into VET does
not influence their labour market outcomes.
A second self-selection of students takes place at the transition from
upper-secondary education to higher education. We argue that the
higher education programmes influence students in such a way that, by
graduation time, those who actually graduate have become much closer
in their knowledge, expertise and practices. Hence focusing on HE
graduates and including control variables for their education, work
experience and employer characteristics should largely account for this
second self-selection.
We find that, one year after graduation, HE graduates with VET work
experience have significantly higher wages of 7% to 19% and need
significantly less time (about two months) to find their first job. How­
ever, we find no significant effect on the likelihood of their doing an
internship after graduation. Although these positive effects from work
experience are very robust to the inclusion of control variables, they
diminish over time: five years after graduation the estimated coefficients
for wages, unemployment and employment position remain positive but
are no longer robustly significant. As for the channel through which VET
work experience operates, we find that the benefits stem from employers
knowing the worker’s productivity level (screening), an increase in
human capital (both general and specific) and employers’ receiving a
signal of higher productivity. Our robustness checks confirm that we are
measuring work experience, that we have representative results for HE
graduates, and that our findings do not suffer from selection bias.
Our results contribute to the literature on VET in three ways. First,
we show empirically that VET work experience from predominantly
dual programmes in a permeable education system has a significant
positive effect on labour market outcomes, even after higher education.
Second, we show that the advantage of VET diminishes only slowly over
time for HE graduates. Third, we identify the channels through which
VET work experience operates as screening, human capital and signal­
ling, not social network. Starting with dual VET is beneficial not only for
early entry into the labour market but also for later entry after higher
education. Although the stigma that VET is a second-best option exists
even in Swiss society, where VET is the majority educational pathway
(Wolter et al., 2014), these results show that VET offers a stronger start
even for eventual HE graduates.
Furthermore, the results underline the importance of having a
permeable education system for complementing the short-term benefits
of dual VET with the long-term benefits of an academic degree. Given
6
In Switzerland the academic diploma is called an ‘academic baccalaureate’
or ‘academic certificate’. In this study we use ‘academic diploma’ for graduates
of academic upper secondary programmes, not only for ease of reading but also
because both are upper secondary diplomas.
7
The 26 regions in Switzerland, called ‘cantons’, are its member states, much
like states in the US.
M.E. Oswald-Egg and U. Renold
that the 2010 Bruges communiqué and Europe 2020 strategy have put
the idea of a permeable education system in the centre of attention in
Europe, one can view VET as a first step in a lifelong learning process
(Cedefop, 2012). Moreover, given that technological and other changes
make lifelong learning a necessity, the permeability of an education
system is particularly important for all countries (Laal & Salamati,
2012).
This article is organised as follows. In Section 2, we discuss work
experience and summarise the literature on causal estimations of work
experience. In Section 3, we present our estimation methods. In Section 4,
we show the results of our analyses and discuss robustness tests on the
validity of our results. In Section 5, we discuss our findings and conclude.
2. Literature review on work experience and hypotheses
Although economists consider work experience when estimating
earnings, no clear economic definition of work experience exists. Most
economic studies treat work experience as a kind of skill (see, e.g. Rosen,
1972; Mincer, 1974; Yamaguchi, 2010). Raelin (1997) defines experi­
ence as the ability to make decisions in complex, novel situations.
Similarly, Gruber (1999) defines work experience as ‘the ability to
repeatedly cope with complex situations that have changing elements’.8
Therefore, work experience is the skill of reacting appropriately to new
and complex circumstances.
Both individuals and the labour market value work experience. For
individuals, having work experience is a good way of differentiating
themselves from other similarly educated job applicants (Tomlinson,
2008). Moreover, work experience is transferable from previous
employment to new employment (Yamaguchi, 2010). The degree of
transferability depends on the extent to which the work experience is
firm-specific or general (Jung & Magrabi, 1991). The more general it is,
the better the work experience can be applied in a new environment
(Burdett, Carrillo-Tudela, & Coles, 2011). Individuals with work expe­
rience also need less training, thereby reducing employers’ training costs
(Thurow, 1975). Hence the labour market values work experience by
paying higher wages and offering better conditions (e.g. Mincer, 1974;
Altonji & Shakotko, 1987; Williams, 1991; Lemieux, 2006; Dustmann &
Meghir, 2005; Heckman, Lochner, & Todd, 2006a), resulting in higher
lifetime earnings for individuals with experience (Mincer, 1974).
The method and means by which individuals obtain work experience
are well known. Individuals gain work experience when they actively solve
a problem (Arrow, 1962). In this process of learning by doing (Dustmann &
Meghir, 2005; Burdett et al., 2011), learning and attaining work experi­
ence are complementary (Rosen, 1972; Mincer, 1974). Therefore, work
experience is not learnt at school but rather best obtained at the workplace.
For example, Mincer (1974) argues that certain skills, such as practical
problem solving, are best learnt at the workplace. Cranmer (2006) finds
that although education institutions attempt to transmit marketable skills,
their success is mixed, and the effort would be better spent on transmitting
real experience. Billett (2008) goes further, arguing that work experience
(‘marketable skills’) can only be acquired at the workplace. From this
perspective, gaining work experience is a way of acquiring marketable
skills that are useful in the labour market (Autor, 2001), skills that are best
acquired – or that can only be acquired – at the workplace.
Because the labour market values workers with work experience,
many students gain this experience during their education. Studies show
that work experience gained during upper-secondary education –
whether through a student job after school or on weekends (e.g. Ruhm,
1997), an internship during holidays (Neyt, Omey, Verhaest, & Baert,
2019) or vocational education and training (VET; e.g. Arum & Shavit,
1995) – has a positive impact on the transition into the labour market
and earnings. Yet, these advantages may come at the cost of worse
educational outcomes (for an overview see Ruhm, 1997), less
8
Authors’ translation from German.
3
Economics of Education Review 80 (2021) 102065
educational attainment (Neyt et al., 2019) or worse long-run labour
market outcomes (e.g. Hampf & Woessmann, 2017).
Students also gain work experience before graduating from higher
education, either before or during their education. Work experience op­
portunities during higher education include student jobs, internships,
sandwich placements (also known as ‘industrial placements’), work-
based higher education that includes both studying and working part-
time, or vocational bachelor and master degrees. Studies analysing the
effects of work experience during higher education on labour market
outcomes find mixed results for HE graduates. For example, for labour
market outcomes, Ehrenberg and Sherman (1987) find no effect of stu­
dent jobs, whereas Baert et al. (2016) find no overall effect of work
experience from a student job but negative effects for certain subgroups.
In contrast, San (1986) and Häkkinen (2006) find positive effects of work
experience from student jobs for HE graduates, although only when HE
graduates did not extend their studies past the official programme length.
Internships, sandwich placements, work-based higher education,
and vocational bachelor or master degrees may have different effects
from general student jobs, because they are often related to the students’
field of study. For internships, Nunley et al. (2016), Margaryan et al.
(2019) and Bolli et al. (2019a) find positive effects of work experience
on labour market outcomes after higher education, but Klein and Weiss
(2011) and Weiss et al. (2014) find no effect. Work experience from
sandwich placements has positive effects on labour market outcomes
(Brooks & Youngson, 2016).
Despite a great deal of evidence on the effect of work experience
from internships on labour market outcomes, we find no such evidence
for the effect of work experience from work-based higher education. The
evidence on vocational bachelor and master degrees is that they increase
earnings (Böckerman, Haapanen, & Jepsen, 2018; Böckerman, Haapa­
nen, & Jepsen, 2019). These overall mixed results for work experience
gained during higher education may arise from the two major draw­
backs of gaining work experience while studying: i) the time it takes
away from studying may result in lower grades, and ii) it may negatively
affect students’ choice to continue with higher education (Stern et al.,
1990; Curtis & Shani, 2002; Bartolj & Polanec, 2018; Neyt et al., 2019).
If students wish to gain work experience before entering higher ed­
ucation, they can enter the labour market for regular employment or
obtain a VET diploma. There is little evidence on the effect of these
sources of work experience on labour market outcomes after higher
education, with the only study on regular employment finding no sig­
nificant effect on labour market outcomes (Weiss et al., 2014). The
disadvantage of these sources of work experience is delayed graduation,
and therefore less time for obtaining returns from their investment in
higher education.
VET’s structured curriculum and recognised diploma differentiate it
from other workplace training, such as internships (Steedman, 2012).
Again, the evidence for the effect of work experience from VET on labour
market outcomes after higher education is mixed. Two studies directly
examine the relation between VET and the labour market outcomes of
HE graduates. Shaw (2012) qualitatively studies the labour market
outcomes after graduation for six female BA (Hons) Educational Studies
graduates in the UK. Three of the six had entered higher education with
an academic diploma; the other three, with a VET diploma. She finds
that having a VET diploma is a positive factor. For Italy, Agarwal et al.
(2019) compare the effect of a VET diploma to that of an academic
diploma after higher education. They find that while a VET diploma
helps HE graduates find jobs more quickly after graduation, it creates
disadvantages for other labour market outcomes such as employment,
wages, working hours and job position.
This contradictory evidence for VET may stem from the variation in
VET programmes relative to workplace training. The OECD (2004) dif­
ferentiates between school-based and work-based (dual) VET by the
proportion of time spent at work. In school-based VET, 75% or more of
the learning takes place at school, while dual VET students spend be­
tween 10 and 75% of their time at school and the rest in training
M.E. Oswald-Egg and U. Renold
workshops or at a company. The workplace training in dual VET allows
VET students to have contact with real clients, experience teamwork
with older co-workers, and apply school-taught theory to practice
(Ryan, 2001; Wolter & Ryan, 2011). In this way, VET students gain work
experience. Dual VET is also more closely linked to the labour market
because of the shared responsibility between the education and
employment systems (Rageth & Renold, 2019). Thus differences among
VET programmes relative to workplace training affect their effects in the
labour market. For example, Bolli et al. (2019b) show that only dual VET
programmes have a positive effect on labour market outcomes.
Tuor and Backes-Gellner (2010) consider different education path­
ways and their impact on labour market outcomes. They find that
following a mixed pathway (vocational upper-secondary education fol­
lowed by academic higher education, or the opposite) increases earnings
after higher education, although they cannot account for student
self-selection into education pathways. They also stress that a permeable
education system is key for this finding, because students must be able to
transition between vocational and academic education. Our study ex­
amines a system in which dual VET is the norm. Because dual VET
provides work experience, and because a permeable education system
with a large share of dual VET in upper-secondary education enables
transitions, we expect VET work experience to generate an advantage for
HE graduates. We therefore hypothesise as follows.
H1: HE graduates with VET work experience will have better labour
market outcomes one year after graduation than HE graduates with an ac­
ademic diploma.
Thus far, no studies have looked at the longer-term impacts of VET
on labour market outcomes after higher education. The evidence on the
long-term effect of work experience shows that the effect of work
experience fades with time due to the depreciation of human capital,
with diminishing returns to work experience (Mincer, 1974; Burda,
2001). Hence, HE graduates with an academic diploma will catch up
over time in terms of labour market outcome. Nevertheless, unless HE
graduates with an academic diploma postpone retirement, the lifetime
earnings of HE graduates with VET work experience will remain larger.
That a reversal in labour market outcomes will occur such that HE
graduates with an academic diploma have better labour market out­
comes is unlikely, because all HE graduates have the same highest ed­
ucation regardless of their upper-secondary education. Thus the
differences in general education should be minimal. Therefore, in
contrast to the literature on long-term effects of VET, we do not expect to
find that VET can become a liability in the long term as a result of fewer
general skills (Forster, Bol, & van de Werfhorst, 2016; Hanushek,
Schwerdt, Woessmann, & Zhang, 2017; Hampf & Woessmann, 2017;
Rözer & Bol, 2019). Furthermore, not all evidence finds a negative effect
of VET. Hanushek et al. (2017) argue that the disadvantage of VET does
not apply to all countries (in this case Switzerland, p. 82). However, for
VET not to be a disadvantage, it has to be mostly dual, contain a great
amount of general education, and be part of a permeable system (Wolter
& Ryan, 2011). We therefore hypothesise as follows:
H2: Five years after graduation from higher education, no significant
difference will remain between the labour market outcomes of HE graduates
with a VET diploma and HE graduates with an academic diploma.
The literature on work experience offers various explanations for how
such experience influences labour market outcomes (Bills, 2003; Weiss
et al., 2014; Margaryan et al., 2019). We examine four of them: social
networks, screening, human capital (specific and general) and signalling.
The social network channel, first defined by Granovetter (1973),
entails building up personal relationships to find employment more
easily in the future. In VET, students have the opportunity to build an
extensive network, through not only their supervisors and colleagues
but also their peers’ co-workers. HE graduates with an academic
diploma do not have that opportunity. The advantage of VET work
experience might be that the network established during VET helps HE
graduates with VET work experience to find a job. We therefore
hypothesise for the social network channel as follows:
4
Economics of Education Review 80 (2021) 102065
H3a: If VET work experience operates through the social network
channel, only HE graduates finding their employment through their social
networks will have better labour market outcomes.
The screening and signalling channels are related, because both entail
efforts at reducing information asymmetry. By screening potential
workers, employers reduce information asymmetry through observation of
the employee at work (Stiglitz, 1975). Before offering employment, em­
ployers might use VET work experience as a screening device to sort the
productive workers from the unproductive ones (Brooks & Youngson,
2016). However, only the VET student’s training company would have that
information. Thus we hypothesise for the screening channel as follows:
H3b: If VET work experience operates through the screening channel,
only HE graduates starting work for a previous employer will have better
labour market outcomes.
According to the human capital channel, a concept dating back to
Becker’s human capital theory (1962; 1964), the labour market pays
workers according to their skills and productivity. Mincer (1974) states
that work experience is a major source of productivity, along with
schooling. Experience also reduces training costs by shortening initial
unproductive time (Thurow, 1975). Some studies argue that VET pro­
vides only specific human capital, which is suitable at the training firm
or a firm in the same industry. However, VET work experience can also
build general human capital, in which marketable skills are transferable
across occupations (Winkelmann, 1996; Rauner & Maclean, 2008;
Shaw, 2012). In dual VET, where the education and employment sys­
tems cooperate in developing national curricula (Rageth & Renold,
2019), VET work experience also entails general human capital.
Therefore, we hypothesise for the human capital channel as follows:
H3c1: If VET work experience operates through the specific human
capital channel, then HE graduates with VET work experience related to their
studies or employment will have better labour market outcomes than HE
graduates with unrelated VET work experience.
H3c2: If VET work experience operates through the general human
capital channel, then HE graduates with VET work experience unrelated to
their studies or employment will have better labour market outcomes than HE
graduates with an academic diploma.
Spence’s (1973) signalling theory states that employers cannot
observe the actual productivity of new workers due to information
asymmetry. To reduce this asymmetry, workers look for ways to signal
their high productivity to employers. Any observable characteristic that
highly productive workers can more easily obtain is suitable for sig­
nalling, such as educational attainment or accumulation of work expe­
rience (Heckman, Stixrud, & Urzua, 2006b). Indeed, Van Belle et al.
(2019) find in a vignette study that work experience signals better work
attitude, larger social network, sense of responsibility, motivation and
maturity to employers. Hence VET work experience may constitute a
signal of higher productivity, because entering higher education calls for
those with a VET diploma to have put in additional effort, i.e. additional
academic education.
We exploit the Swiss education system’s having both universities and
UASs. Although the State Secretariat for Education, Research and
Innovation promotes the two higher education types as equivalent, they
differ in their curricula, so that employers perceive UAS graduates as
having more hands-on experience than university graduates. Thus when
VET has the same effect for graduates of either type of higher education,
signalling is the best explanation. We therefore hypothesise for the
signalling channel as follows:
H3d: If VET work experience operates through the signalling channel,
then the effect of VET work experience on the labour market outcomes for
university graduates and UAS graduates will be the same.
Thus far, there is no evidence on the channel through which VET
work experience operates. While studies on work experience from in­
ternships during higher education predominantly find signalling to be
the operational channel (e.g. Weiss et al., 2014; Nunley et al., 2016),
evidence for the general human capital channel also exists (Bolli et al.,
2019a).
M.E. Oswald-Egg and U. Renold
3. Methodology
In this section we describe our dataset, define the variables for the
analysis and explain the estimation methods.
3.1. Data
Our data stem from the Swiss graduate survey conducted by the
Swiss Federal Statistical Office (SFSO) every second year on all gradu­
ates from formal institutions of higher education. The SFSO survey
contains detailed information on employment and education during
studies, and for both one and five years after graduation. Our dataset
consists of pooled cross-sectional data from the one-year post-gradua­
tion survey waves in 2011, 2013, 2015 and 2017, and the five-year post-
graduation survey waves in 2015 and 2017.9
We limit our analysis to HE graduates from fields of study in which
work experience is not largely included, nor an internship required, after
graduation. We therefore drop all HE graduates with a teaching, medi­
cal, law or doctoral degree. Next, to have HE graduates with uninter­
rupted study paths – thereby avoiding unobserved work experience – we
restrict our sample to HE graduates not older than 30. Further, they must
have completed their entry diploma for higher education (i.e. academic
or vocational diploma) within seven years before higher education
graduation. This time span gives them five years for studying for both
bachelor and master degrees, an additional year for repeating a semes­
ter, and another year for fulfilling potential entry requirements. Finally,
we focus on HE graduates from the Swiss education system who tran­
sitioned directly into the labour market. Therefore, we drop all HE
graduates who have a non-Swiss educational background, acquired a
higher education degree abroad, continue their higher education
studies, are self-employed, or do not seek employment due to family,
health or other issues.
We consider a set of dependent variables because, as Nunley et al.
(2016) argue, work experience might operate through different chan­
nels, thereby influencing the labour market outcomes differently at
different stages in the recruitment process. One year after graduation,
only a handful of HE graduates remain unemployed. Therefore, we
analyse the impact of work experience on wages, search time and in­
ternships (within one year after graduation). To analyse longer-term
consequences, we consider wages and employment positions five years
after graduation, as well as whether HE graduates had any unemploy­
ment spells during that period.
Wage is important because it measures the labour market value of an
employee. We define gross wage – expressed in full-time employment –
as comprising contract wage, wage from additional hours, and bonuses.
Time of job search stands for a period of wage absence and insecurity:
the shorter, the better. We construct the search time variable by accu­
mulating all search time before and after graduation in months
(censored at 24 months), including HE graduates who were offered a job
at zero search time and therefore did not seek employment.
We capture work place security and forgone wages by looking at
internships after graduation, at the employment position and at
9
The average response rate for the one-year post-graduation surveys is 60%
of all HE graduates. For the five-year post-graduation surveys, 70% of all
contacted HE graduates taking part in the first survey and agreeing to take part
in a future survey responded. We cannot rule out estimation bias from non-
response, such as HE graduates who did not find employment not taking part
in the survey. This issue potentially influences the size of the coefficients in our
estimations, under- or overestimating the actual effect. For the overall findings,
however, where we compare HE graduates with a VET diploma to HE graduates
with an academic diploma, the potential bias becomes a problem only if the
survey participation of unemployed HE graduates with a VET diploma sys­
tematically differs from that of unemployed HE graduates with an academic
diploma. We see no reason why unemployed HE graduates should behave
differently from one another, no matter what their upper-secondary education.
5
Economics of Education Review 80 (2021) 102065
unemployment spells. As internships pay less – if they have a wage at all
– than regular employment, HE graduates entering the labour market via
an internship forgo wages (generation internship; for more information
see Economist, 2014; Schmidlin & Witmer, 2007; Cerulli-Harms, 2017).
Our variable for employment position has five levels: apprentice,
non-managerial employee, lower management employee, middle man­
agement employee and higher management employee. Unemployment
spells indicate whether HE graduates were unemployed during the five
years after graduation.
Given the permeability of the Swiss education system, Swiss HE
graduates can choose between two main upper-secondary pathways
towards a higher education degree (SERI, 2020; Hoffman & Schwartz,
2015). They decide between obtaining an upper-secondary academic
diploma and an upper-secondary VET diploma. The four-year academic
diploma qualifies its graduates for direct entry into a university or fed­
eral institute of technology (Uni). It also gives them access – albeit
indirectly, through a working year – to UASs.
The VET diploma, which takes three to four years to complete,
qualifies its graduates for the labour market or a UAS when attending
school an additional day per week. To enter a Uni, VET graduates must
have an additional year of academic education. As long as students
fulfil these entry requirements, they are free to attend any university or
UAS.
One main difference between the two education pathways at the
upper-secondary level is that the academic diploma involves a
completely theoretical education, whereas the VET diploma includes
both theoretical education and practical training. In Switzerland, about
two-thirds of students completing compulsory education enrol in VET.
Most enter a dual VET programme (SERI, 2020), in which the practical
training is performed at a workplace. Therefore, graduates with a VET
diploma have not only an academic education but also work experience
gained during their education. In contrast, graduates with an academic
diploma have no in school work experience. This difference in work
experience is precisely what we analyse: our variable of interest is
whether the HE graduate has obtained an academic diploma or a VET
diploma10 in upper-secondary education.
We can account and control for observable differences in personal
characteristics, education, further education, experience and current
employer characteristics, all of which may affect labour market out­
comes (Altonji & Blank, 1999). For personal characteristics we consider
age, gender (male/female), parents’ highest education (socio-economic
background), marital status, having children, region of residence and
region of origin. For education we control for type of institution
(Uni/UAS), level of education (BA/MA), educational institutions (20),11
subject (10),12 number of semesters, final grade and cohort.
For further education we have information on continuing studies (e.
g. doctorate), additional studies (i.e. non-university education), and
10
To enter higher education, VET students may attend school for an addi­
tional day each week instead of being at the workplace. About 15% of VET
students choose that option. Thus HE graduates with a VET diploma have at
least the additional day at vocational baccalaureate school when at UAS or even
the additional academic year (university aptitude test) when at university. As
not all VET students are able to obtain these additional requirements, we
encounter sorting of VET students into the higher education programmes.
However, because our analyses focus only on comparing HE graduates with a
VET diploma to HE graduates with an academic diploma, this sorting among
VET students does not bias our estimations.
11
The 20 education institutions are universities from Basel, Bern, Fribourg,
Geneva, Lausanne, Lucerne, Lugano, Neuchatel, St. Gallen and Zurich as well as
the Swiss Federal Institutes of Technology in Lausanne and Zurich. The UASs
are from Bern, Central Switzerland, Eastern Switzerland, Kalaidos, North­
western Switzerland, Southern Switzerland, Western Switzerland and Zurich.
12
The 10 education subjects are the humanities, the arts, educational sciences,
economics, natural sciences, medical sciences, health, engineering, agricultural
sciences and ‘interdisciplinary’.
M.E. Oswald-Egg and U. Renold
Fig. 1. Variation of VET enrolment rate by region and within region for
different cohorts. Note: The figure shows variation in enrolment rates in VET
both across regions and within regions, because the enrolment rate in a region
fluctuates from year to year.
continuing education (e.g. course or training). For life experience we
differentiate between kinds of work experience (i.e. work experience
gained before higher education, student internship, student job and
work-based higher education) and other experience (e.g. a year at
another university). For employer characteristics we use whether the
company is from the public sector, size according to full-time employees,
industry (21 on NOGA-1),13 and the labour market region of the
employer (17).14, 15
However, HE graduates do not randomly pick their education
pathway but rather have unobservable selection criteria. Thus we face
an endogeneity problem, or the risk of an omitted variable confounding
with the labour market outcomes and the decision to obtain a VET
diploma. For example motivation to work could be higher for people
choosing the VET diploma, in turn leading to higher labour market
outcomes even had they not chosen the VET pathway. Not accounting
for such issues biases the estimates. As instrumental variables constitute
one way of solving this problem, we use a variable correlating with the
decision to obtain a VET diploma but not with the labour market out­
comes of HE graduates (Angrist, Imbens, & Rubin, 1996; Angrist &
13
The 21 industries are agriculture, forestry and fishing; mining and quar­
rying; manufacturing; electricity, gas, steam and air conditioning supply; water
supply; sewerage, waste management and remediation activities; construction;
wholesale and retail trade; repair of motor vehicles and motorcycles; trans­
portation and storage; accommodation and food service activities; information
and communication; financial and insurance activities; real estate activities;
professional, scientific, and technical activities; administrative and support
service activities; public administration and defence; compulsory social secu­
rity; education; human health and social work activities; arts, entertainment
and recreation; other service activities; activities of households as employers;
undifferentiated goods- and services-producing activities of households for own
use (no data); and activities of extraterritorial organisations and bodies.
14
The 17 labour market regions are Aarau-Olten, abroad, Basel, Bellinzona,
Bern, Biel, Chur, Geneva, Fribourg, Lausanne, Lugano, Luzern, Neuchatel, Sion,
St. Gallen, Winterthur-Schaffhausen and Zurich.
15
As most of the controls are measured after the choice of upper-secondary
pathway, they might be an outcome of the pathway choice, that is, bad con­
trols. This concern particularly affects type of institution, subject, company size
and industry. However, the choice of upper-secondary pathway does not restrict
students to either a type of institution or a subject. The permeability of the
Swiss education system allows switching at relatively low costs, if any at all.
Our focus is on HE graduates, whose higher education will weigh much more in
the choice of company and industry than their upper-secondary education.
Therefore, we argue that these controls are not bad at all.
6
Economics of Education Review 80 (2021) 102065
Krueger, 2001). However, this method provides causality only for the
group of HE graduates who would have changed their choice had the
instrument changed (local average treatment effect, LATE). Therefore,
the external validity of the results is limited, an issue that we discuss
more thoroughly in the conclusion.
Some studies use regional borders as the instrumental variable for
looking at the effect of an intervention (Bolli & Hof, 2018; West &
Woessmann, 2010; Frölich & Lechner, 2010; Card & Krueger, 1994).
Following Bolli and Hof (2018), we use the VET enrolment rates of the
HE graduates’ pre-study region (origin) at the time they were in their
second year of lower-secondary education, when they had to decide
their future education pathway. Fig. 1 shows the variation in VET
enrolment rates among regions and within regions for individuals of
different cohorts. The regional VET enrolment rate at the time of edu­
cation decision correlates with the decision to enrol in VET. The
first-stage coefficients are significant and robust (see appendix A.2). This
variable also fulfils the exclusion restriction, because students are mo­
bile within Switzerland after their graduation from higher education
(Oggenfuss & Wolter, 2018), so that their labour market outcomes do
not depend on the VET enrolment rate in their region of origin.
To account for the second self-selection of students from upper-
secondary education into higher education, we focus on HE graduates
and include a large set of control variables on their education, work
experience and employer characteristics. The focus on HE graduates is
important because in general only the more able VET students, who
obtained a vocational diploma or even the additional year of academic
education will continue with higher education. In contrast, most stu­
dents with an academic diploma will continue with higher education.
However, Bolli, Rageth, and Renold (2019c) show that although both
pathways in upper-secondary education have high-ability students, the
academic pathway has more of the students with higher cognitive
abilities, so that the actual differences in ability between students with a
VET diploma or an academic diploma entering higher education remain
unclear. Given that in Switzerland everyone with a diploma (academic
or vocational) is allowed to enter higher education without further entry
examinations, the sorting of students takes place in the first year of
higher education. We argue that this sorting of more able students and
the following higher education studies influence students in such a way
that those who graduate have become very similar in their abilities.
Hence the inclusion of the extensive set of control variables should ac­
count for the remaining differences.
To identify the channel through which work experience from VET
operates, we use three additional variables. The first indicates whether
or not the VET diploma is related to the educational subject, to
employment, or to both.16 The second variable is on HE graduates’
16
The variable VET diploma – related to education subject, employment, or
both – is a binary variable that we generate from four questions on the HE
graduates’ survey. The first question asks HE graduates to rate (on a five-point
Likert scale) whether there is a connection between their VET diploma and
higher education studies. We attribute values of 4 and 5 to the VET diploma’s
being related to the higher education study generating a one in our binary
variable; otherwise the variable is zero. The second question asks HE graduates
whether their current employment is related to the university education, pre­
vious employment or both. The options are i) yes, in relation to both the uni­
versity education and previous employment; ii) yes, in relation to the university
education; iii) yes, in relation to the previous employment; and iv) no to uni­
versity education or previous employment. Our binary variable is already one
for options i) and ii) from the previous question. Thus the variable needs to
become one for option iii), but only when the previous employment was (most
probably) VET. The answers to the following two questions make our variable
more exact. The third and fourth questions ask graduates about their engage­
ment in paid employment (related or unrelated to their studies) during their
studies. We change our binary variable to one only if the answers to the third
and fourth questions are ‘never’.
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

finding employment through their social networks (family and previous
employer).17 The third variable captures HE graduates’ starting
employment at a previous employer.
3.2. Estimation method
To estimate the effect of work experience from VET on the labour
market outcomes, we use explanatory analysis, exploiting our rich set of
control variables. Thus we tackle endogeneity by including controls for
presumably all differences between HE graduates with VET diplomas and
those with academic diplomas. For wages one year and five years after
graduation, we assume a log-level specification following the Mincer
earnings function (Mincer, 1974) and use a tobit model with
left-censoring at zero. We also assume a log-level function in line with
survival analysis for search time until first employment, using the Cox
proportional hazard model for estimation. The regression equation reads:
log(yi + 1) = β0 + β1 VETi + β2 Xi + ϵi
The yi stands for a labour market outcome of a graduate, i. β0 is a
constant, β1 the coefficient of interest as it tells us the effect of experi­
ence from VET, β2 is a vector of coefficients for the control variables, and
ϵi stands for the individual-specific error term. VETi indicates the edu­
cation pathway taken in upper-secondary education, and Xi is the vector
of control variables, which include education of father, education of
mother, region of residence, institution, subject of study, cohort, com­
pany size, industry and labour market region as fixed effects. We include
these fixed effects to account for the possibility that participation in VET
may be induced by background factors or students’ foreseeing the de­
mand for specific skills in key industries or in their local areas.
For the other labour market outcome variables – internship during
the first year after graduation, unemployment spells during the five
years after graduation, and the employment position five years after
graduation – the underlying function is level-level:
yi = β0 + β1 VETi + β2 Xi + ϵi
However, due to the different ranges of the dependent variables, we
use a probit model for internship and unemployment, and an ordered
probit model for the employment position. All models are estimated
with robust standard errors.
Although our set of controls is very rich, we cannot argue that we
capture all differences. Therefore, to tackle unobserved heterogeneity
from confounding factors, we use an instrumental variable (IV)
approach as a second specification. This estimation consists of two
stages. The first stage (FS) estimates having a VET diploma ( VET
̂ i ) from
the enrolment rate into VET of the region of origin (regVET), and the
second stage (SS) uses the estimation for the VET diploma to analyse its
effect on the HE graduates’ labour market outcomes. to further reduce
heterogeneity (Xi ), we include our control variables in both regressions.
The γ are coefficients in the first stage, and ηi the individual-specific
error term of the first stage. The error terms are robust.
̂ i = γ0 + γ 1 regVET + γ2 Xi + ηi
FS : VET
̂ i + β 2 Xi + ϵ i
SS : yi = β0 + β1 VET
As VET is a binary variable, we use a probit specification for the first
stage and various models for the second stages. We use the OLS estimator
with left-side truncation at zero for wages one year and five years after
graduation and use interval regression with left-side truncation at zero for
the search time (Bartus & Roodman, 2014). The estimation method for
internship during the first year after graduation and unemployment spells
is probit, and the method for employment positions is ordered probit.
Conditional mixed-process models (command ‘cmp’ in Stata; Roodman,
2011) are optimal for modelling the two stages with different estimators.
For these estimations, we also use robust standard errors.18
We perform subsample analyses by using the previous estimation
methods to identify the channel through which VET work experience
operates. However, we only perform this analysis for the dependent vari­
ables for which we initially found significant results – search time and wage
one year after graduation. We examine the social network channel by
comparing HE graduates with a VET diploma who find their job through
their social networks with those who do not. For screening, we compare HE
graduates with a VET diploma who work for a previous employer to those
who do not. To identify the specific human capital channel, we compare HE
graduates with work- or study-related VET and HE graduates without it. To
(1)
examine whether VET has general human capital component, we compare
HE graduates with a VET diploma unrelated to their studies or work to HE
graduates with an academic diploma, restricting our initial sample to only
those graduates and estimating the standard models. To examine the sig­
nalling channel, we analyse the differences between HE graduates from
universities and those from UASs with a VET diploma.
4. Results
Subsection 4.1 briefly describes our variables in the dataset. In
subsection 4.2 we present results on the effect of VET work experience
on the labour market entry outcomes of HE graduates. In subsection 4.3
we examine whether our results are a short-term effect or if they persist
four years later. In subsection 4.4 we examine the channel through
which VET work experience operates. In subsection 4.5 we present
robustness checks to support that our results are measuring work
experience, are representative for Swiss HE graduates, and are exam­
(2)
ining comparable groups with each other.
4.1. Descriptive statistics
In our dataset 38% of HE graduates have a VET diploma (5341,
compared to 8558 HE graduates with an academic diploma). The cor­
relation matrices in Table A.1 and Table A.2 of Appendix A.1 show how
the variables in our dataset correlate with one another for both one year
and five years, respectively, after graduation. For the important relation
between having a VET diploma and the dependent variables, we find
that VET is positively correlated with wage and employment position
but negatively correlated with search time, an internship after gradua­
tion and unemployment spells.
We focus on the differences between HE graduates with an upper-
secondary academic diploma (ACC) and an upper-secondary VET
diploma (VET). We report the descriptive statistics separately for those
groups. The left side of Fig. 2 displays the dependent variables for HE
(3)
graduates with ACC and VET one year after graduation. We observe that
HE graduates with VET earn on average roughly CHF 78,700 (∼ 73,300
(4)
Euro) in the first year after graduation, which is about CHF 9000 (∼
8400 Euro) more than HE graduates with ACC. On average, HE gradu­
ates with VET also search three-fourths of a month less for employment,
i.e. searching for four months while HE graduates with ACC search for

17
We construct this variable from a survey question asking graduates which of
their search strategies was decisive for finding employment. The two answer
options were i) seek employment through people I have met in connection with
a student job (or internship or project work) and ii) seek employment with the
help of personal relationships (relatives, friends, acquaintances, fellow stu­
dents). From this question we generate a binary variable. Employment through
social network is one if the answers were i) or ii) and zero otherwise.
18
None of the estimations includes survey-sampling weights because, ac­
cording to Solon, Haider, and Wooldridge (2015), including them in causal
effects estimations is not essential and in some cases even harmful. However,
they recommend checking whether estimates with and without weighting
differ. If they differ, that difference indicates heteroscedasticity or endogenous
sampling problems. Therefore, we perform the estimations of our main findings
with and without the weights. However, the results show no difference.

7
M.E. Oswald-Egg and U. Renold
Fig. 2. Descriptives of dependent variables by upper-secondary education pathway.
4.74 months. About one in every 10 HE graduate with VET starts the
career with an internship, whereas twice as many HE graduates with
ACC (one in five) do so. The regional enrolment rate in VET is higher for
HE graduates with VET (75%) than for those with ACC (70%).
Concerning the control variables, on average 34% of HE graduates
with VET are female, compared to 60% of HE graduates with ACC. HE
graduates are on average 26 years old, and 6% are married. The parents
of HE graduates with VET have slightly more vocational qualification
than those of HE graduates with ACC. HE graduates with ACC study at
UASs 40% of the time, compared to 96% of HE graduates with VET. Just
as they choose different institutions, HE graduates with VET or ACC also
choose different subjects. The average grade of HE graduates with ACC
(5.14) is slightly higher than for those with VET (5). Both groups of
graduates have work experience from pre-study jobs (around 90%) and
student jobs (roughly 87%), although HE graduates with ACC have more
internship experience (59% vs. 26%), a finding that is not surprising,
given that an internship is a UAS admission requirement for HE gradu­
ates with ACC. While HE graduates with ACC are more likely to work for
a public employer (36% vs. 27%), company size, industry and employer
8
Economics of Education Review 80 (2021) 102065
region are comparable.
In 71% of cases, HE graduates with VET obtained their VET diploma in
an occupation related to their studies, current employment or both. In
finding employment, 22% of HE graduates with ACC rely on their social
network, while only 19% of HE graduates with VET. Furthermore, 15% of
HE graduates with VET return to a previous employer, whereas 8% of HE
graduates with ACC do. Table A.3 in Appendix A.1 contains the details.
The right side of Fig. 2 shows the dependent variables for HE grad­
uates with VET or ACC five years after graduation. We observe that
wages of HE graduates with VET have increased to CHF 90,400 (∼
84,200 Euro) and are still higher by CHF 7000 (∼ 6500 Euro) than those
of HE graduates with ACC wages. HE graduates with VET were about
one-fourth less likely to be unemployed during the post-graduation
period, with only about every seventh unemployed, compared to
every fifth HE graduate with ACC. HE graduates with VET have slightly
higher employment positions (2.6 vs 2.5). In this dataset, as in the pre­
ceding one, the regional enrolment rate in VET is 71% for HE graduates
with ACC and 75% for HE graduates with VET.
For the difference between HE graduates with VET or ACC in the
Fig. 3. Estimation results for labour market out­
comes one year after graduation from higher ed­
ucation. Note: Panel A shows the effect of VET on
wages one year after graduation from higher ed­
ucation. The blue dots show the results from the
tobit estimation; the pink dots, the results from the
IV estimation. Panel B shows the effect of VET on
search time until the first job after graduation
from higher education. The blue dots show the
hazard ratios from the Cox survival estimation; the
pink dots, the coefficients from the IV estimation.
Panel C shows the effect of VET on internship after
graduation from higher education. The blue dots
show the marginal effects from the probit esti­
mation; the pink dots, the marginal effects from
the IV estimation. The horizontal lines in all panels
indicate the point estimates with robust 95%
confidence intervals. Tables A.5- A.7 in
Appendix A.2 show the estimated coefficients.
The tables display all estimates, whereas the figure
does not display the estimates of the fixed effects
for fathers’ highest education, mothers’ highest
education, region of residence, institution, sub­
ject, cohort, size of employer, industry, or labour
market region. To support the stability of the co­
efficients, these tables also report the estimates
without control variables and with region of origin
fixed effects. (For interpretation of the references
to colour in this figure legend, the reader is
referred to the web version of this article.)
M.E. Oswald-Egg and U. Renold
control variables, we find the same patterns as in the preceding dataset
for gender, parents’ education, institution, subject, education level and
employers characteristics. They have the same average age (29), are
married equally often (21%), are equally likely to have children (13%),
and attain further education at the same rate (10% additional studies,
12% continuing education, and 12% continuing studies). Table A.4 in
Appendix A.1 displays all the details.
4.2. The effect of VET work experience on higher education graduates’
outcomes for labour market entry
We start our analysis by estimating the effect of VET work experience
on HE graduates wages, search time until first employment, and in­
ternships one year after graduation. Fig. 3 shows the results for all three
labour market outcomes.
Panel A shows the estimation results for wages. Having a VET
diploma in the tobit (multivariate) estimation model increases wages by
6.9% for an average HE graduate (average treatment effect, ATE). The
results for the estimates from the IV model suggest also that having a
VET diploma increases wages for HE graduates. The wage increase is
18.9% for HE graduates, who are likely to be affected by the instru­
mented variable (local average treatment effect, LATE). To test the
plausibility of our estimation, we analyse the other variables. As ex­
pected, age, master degree and grade have significant positive effects on
wages, whereas being female has a significant negative effect. We also
analyse the robustness of our results by looking in the tobit regression
and the IV regression at the effect without controls.
For the IV regression we further check whether our results still hold
when we include region of origin fixed effects. This variable accounts for
the difference in the enrolment rates into VET and ACC between regions,
a difference that could lead to differences in wages and invalidate our
instrument. Table A.5 in Appendix A.2 shows the estimated coefficients.
For wage the coefficient for the models without control are larger, but
they remain significantly positive. Including region-of-origin fixed ef­
fects does not change the results.
Panel B shows the estimation results for search time of HE graduates
9
Economics of Education Review 80 (2021) 102065
Fig. 4. Estimation results for labour market
outcomes five years after graduation from higher
education, Note: Panel A shows the effect of VET
on wages five years after graduation from higher
education. The blue dots show the coefficients
from the tobit estimation; the pink dots, the co­
efficients from the IV estimation. Panel B shows
the effect of VET on unemployment spells within
five years after graduation from higher educa­
tion. The blue dots show the marginal effects
from the probit estimation; the pink dots, the
marginal effects from the IV estimation. Panel C
shows the effect of VET on employment position
five years after graduation from higher educa­
tion. The blue dots show the coefficients from the
ordered probit estimation; the pink dots, the co­
efficients from the IV estimation. The horizontal
lines in all panels indicate the point estimates
with robust 95% confidence intervals.
Tables A.8- A.10 in A.2 show the estimated co­
efficients. The tables display all estimates,
whereas the figure does not display the estimates
of the fixed effects for fathers’ highest education,
mothers’ highest education, region of residence,
institution, subject, cohort, size of employer, in­
dustry, or labour market region. To support the
stability of the coefficients, these tables also
report the estimates without control variables
and with region of origin fixed effects. (For
interpretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article.)
until first employment after higher education. The resulting hazard ratio
from the Cox survival estimation model (multivariate) indicates that HE
graduates with a VET diploma take less time to find employment (hazard
ratio above 1). However, the result is not significant. The coefficient for
the IV model is significantly negative, indicating that HE graduates with
a VET diploma take about two months’ less time to find employment. In
sum, while the ATE for VET is not significant, the LATE is.
Comparing the size of the effect with the other variables, we find that
while the Hazard ratio is comparable, the IV estimation is much larger
than the effect of the other variables including the other work experi­
ence variables. Looking at the robustness of the finding, we see that the
coefficients without controls have the same sign as those with controls
but that only the Cox model remains significant (see all results in
Table A.6 in appendix A.2). Including the region-of-origin fixed effects
does not influence the IV coefficients. Thus we conclude that VET work
experience significantly reduces search time for HE graduates.
Panel C displays the marginal effects for doing an internship after
graduation from higher education. In the probit model (multivariate), the
marginal effect of a VET diploma is -0.05, meaning that the probability of
doing an internship decreases by five percentage points when the HE
graduate has a VET diploma (ATE). In the IV model, although we find the
same marginal effect of five percentage points, this effect is not significant
(LATE). The size of the effect has a similar magnitude to that of the other
effects. Excluding the controls increases the size of the coefficients and is
significant for IV (all estimates in Table A.7 in appendix A.2). However,
including the region-of-origin fixed effects does not change the IV co­
efficients. Therefore, we do not find a significant effect of VET on reducing
internship experience after graduation from higher education.
4.3. The stability of VET work experience over time
We now examine whether the benefit of VET on HE graduates’ labour
market outcomes persists after four years, or whether the initial advan­
tage turns into a disadvantage. Fig. 4 shows the results for wages five
years after graduation, unemployment spells and employment positions.
Panel A shows the results for wages five years after graduation from
M.E. Oswald-Egg and U. Renold
higher education. The coefficients for VET are positive but smaller than the
previous estimate (2.7% vs. 11.8%). While the coefficient of the IV model
remains significant, that of the tobit model does not. However, the IV model
no longer works because the first stage is insignificant. The estimated co­
efficients are plausible, as age and master increase wages, but female de­
creases them. As excluding the control variables results in significant
coefficients, the results are not robust. Introducing region-of-origin fixed
effects into the IV does not influence the results, and the first stages are also
insignificant (see Table A.8 in appendix A.2). Therefore, we do not find a
robust effect of VET on wages five years after higher education.
Panel B displays the result for unemployment. We find no significant
evidence for the influence of a VET diploma on HE graduates being
unemployed during the first five years after graduation. Both estimation
methods find a small positive effect. However, excluding the control
variables changes the sign, so that the effect becomes significantly
negative (all results appear in Table A.9 in appendix A.2). As for wage,
the first stage becomes insignificant when we include controls and re­
gion dummies in the IV estimation. Therefore, we find no evidence that
VET work experience prevents unemployment after graduation.
Finally, Panel C shows the evidence for the effect of VET work
experience on employment positions. Both coefficients for VET are
positive but not significant. As in the two previous labour market out­
comes, the results are not robust. When we exclude control variables, the
results become significant, and when we add regional dummies they
change sign (see results in Table A.10 in appendix A.2). Therefore, we
find no robust evidence that VET work experience significantly increases
employment positions five years after graduation.
4.4. The channels through which work experience from VET operates
After finding that VET work experience has significant influence on
short-term labour market outcomes, we extend our analysis to examine the
four theoretical channels through which VET work experience may be an
advantage (all results appear in Table 1). For these analyses, we restrict our
sample to HE graduates with a VET diploma and assume that the findings
are transferable to the comparison of HE graduates with ACC and VET.
To analyse the social network channel, we compare HE graduates
who found employment through their social networks with HE gradu­
ates who did not. We find that social network increases search time, but
does not influence wage. Hence the VET diploma does not operate
through the social network channel.
The analysis for the screening channel is similar to the one for the
social network channel. We compare the effect of VET work experience
when HE graduates find employment with a new employer to the effect of
VET when HE graduates return to a previous employer. We cannot
examine search time, because all HE graduates returning to their previous
employer have zero search time. For wage, the coefficient of screening is
significantly positive. Therefore, the VET diploma works for HE gradu­
ates through the screening channel for wages, increasing them by 4.7%.
Comparing HE graduates who have obtained their VET diploma in a
field related to their employment or studies with HE graduates with an
unrelated VET diploma, we find that HE graduates with ‘related’ VET have
significantly higher wages (4.3%). This finding indicates a specific human
capital component to VET, a component that generates generates benefits
for wages but does not significantly affect search time. The results for
comparing HE graduates with unrelated VET to HE graduates with ACC
shows that HE graduates with unrelated VET still have an advantage over
HE graduates with ACC in wages that are 4% higher and in reduced search
time. Therefore, VET also has a general human capital component.
As to whether VET work experience operates through the signalling
channel, we compare the effect of VET work experience of HE graduates
of Unis with HE graduates of UASs. We find no effect for wages or search
time. Therefore, VET signals higher productivity to employers.
In sum, we find that the advantages of a VET diploma for HE grad­
uates operates through the screening, human capital and signalling
channels, transmitting both specific and general human capital.
10
Economics of Education Review 80 (2021) 102065
Table 1
Regression table to identify channel.
Wage Search time
N Coef. SE HR SE
Social network 4706/4706 -0.008 0.01 0.922** 0.03
Screening 5928/4995 0.047*** 0.01 - -
Specific HC 5975/5042 0.043*** 0.01 0.994 0.03
General HC 10727/9760 0.040*** 0.01 1.069** 0.03
Signalling 5975/5042 -0.031 0.02 0.958 0.07
Controls yes yes
Note: The table shows the regression coefficients (Coef.) from the tobit regres­
sion for wage and the hazard ratios (HR) for the Cox regression for search time. N
indicates the number of observations for wage/search time. The first four esti­
mations include all the previous controls, and the standard errors (SE) are
robust. The last estimation (signalling) includes all the previous controls except
the education institution, and the standard errors are robust. ***,** and * denote
significance at the 1%, 5% and 10% level, respectively.
4.5. Robustness checks
As we make certain assumptions in our analyses, we now test these
assumptions with empirical data.
4.5.1. Are we really measuring work experience?
Since Mincer (1974) economic scholars have used work experience
and its quadratic form for the estimation of wages (also called earnings).
However, scholars still debate how to approximate work experience.
Some studies use potential work experience by taking a person’s age and
subtracting the number of school years and the entry age for school (e.g.
Altonji & Blank, 1999). Others estimate actual work experience by
looking at the work history of individuals (e.g. Light & Ureta, 1995).
These procedures, however, generate an attenuation bias due to their
crudeness (Gullason, 1990; Altonji & Blank, 1999; Blau & Kahn, 2013):
they neglect the quality of work experience and the possibility of
obtaining work experience during schooling. Economic scholars have
developed new methods for measuring actual work experience (Light &
Ureta, 1995; Regan & Oaxaca, 2009; Zveglich, Rodgers, & Lavina, 2019)
and for capturing its quality (e.g. Gullason, 1990).
Nonetheless, one difficulty in measuring work experience is that
more than one ideal type of work experience exists (Blackwell, Bowes,
Harvey, Hesketh, & Knight, 2001). Quiñones, Ford, and Teachout
(1995) avoid this difficulty by establishing a conceptual framework for
classifying work experience along on two axes: measurement mode and
level of specificity. Three elements contribute to their measurement
mode: number of times that something (e.g. a task or job) was done, time
spent doing it and the respective difficulty or complexity of doing it.
Their level of specificity axis captures gaining work experience at
different levels: task, job or organisation. Together, the two axes
constitute a nine-cell matrix (p. 892).
In our focus on VET work experience, differentiation at the task level
is sufficient because no job or organisation change is involved in VET.
We focus on the number of times the task was performed, time spent on
the task and the difficulty or complexity of the task. When we apply
Quiñones et al.’s (1995) framework to VET, we find that students
perform the same tasks over and over in a workplace, and that the time
spent on the tasks will differ according to the required days at school in
the occupational curriculum. The complexity of the task will also differ,
especially in dual VET, because clients’ varying needs lead to a variation
in the complexity. For example, although a hairdressing apprentice will
cut hair almost daily during dual VET, the time he or she spends on the
haircut will differ according to the length of the hair, and the difficulty
will depend on the hairstyle the client wants. Likewise, a salesperson in a
store is responsible for handling cash at checkout but also for placing
products on shelves, knowing the products well enough to give clients
advice, and understanding the entire flow of warehouse goods.
Therefore, in Quiñones et al.’s (1995) framework, HE graduates with
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table 2 and about the same amount of internship than our sample. Lower wages
Regression table on work experience. are expected because students remain in the dataset. A longer search time
Wage IVWage is unexpected, because students should not have had any. Finally, the
proportion of students with a VET diploma in the dataset is lower (23%)
Coef. SE Coef. SE
than that in our sample (38%), indicating that HE graduates with VET are
First stage: VET indeed over-represented in our sample. The regional enrolment rate in
Regional enrolment rate in VET 0.867*** (0.28)
Second stage: Wage
VET is about the same in both the dataset and our sample.
VET 0.065*** (0.01) -0.396*** (0.02) For personal characteristics, the students in the dataset and our HE
Pre-study job -0.012 (0.01) -0.029** (0.01) graduates are comparable. The students in the dataset are only slightly
Related pre-study job 0.027*** (0.01) 0.047*** (0.01) older (27 rather than 26) and more likely to be married and have chil­
Student internship 0.029*** (0.01) -0.022** (0.01)
dren. In the education characteristics we find no differences in the
Student job -0.038*** (0.01) -0.037*** (0.01)
Related student job 0.050*** (0.01) 0.063*** (0.01) proportion of master students, the average final grade or the number of
Work-based high. edu. 0.049*** (0.01) 0.074*** (0.01) semesters. However, our sample contains an over-representation of HE
Observations 13,490 13,490 graduates from UASs. The largest differences in the subjects are for
Pseudo R2 0.659 economics and agricultural sciences, for both of which we have an over-
Note: The table shows the regression coefficients from the tobit regression and IV representation. Finally, practically no differences exist between the
regression for wage. The estimations from the first and second stages include all dataset and our sample for work experience and its source.
the previous controls, and the standard errors are robust. ***, ** and * denote This comparison shows that our sample contains a higher rate of HE
significance at the 1%, 5% and 10% level, respectively. graduates with VET and from UASs. As these two features are critical for our
analysis, they might be driving our results. To rule out any special influence
a VET diploma have work experience. The question then is whether we of these features, we estimate a Heckit – a two-stage Heckman correction
really measure work experience or something else connected to the VET model (Heckman, 1979). As an exclusion restriction we use two instruments:
diploma. One possible way of checking whether we actually measure continued study and searched employment. Table A.12 in Appendix A.3
work experience is to compare our coefficients from VET to the other shows the results: the coefficient for VET is still significantly positive for
work experience coefficients we have in our estimations. We can wage. Hence the selection in our sample does not drive our results.
differentiate six sources of work experience, four that are related to
students’ field of study and two that are not. Internships and work done 4.5.3. Are we comparing apples and oranges?
by part-time students (work-based higher education) might be included Another assumption we make is that HE graduates with VET or ACC
in the curriculum, and we can observe which pre-study work experience are comparable, because having the same degree from the same insti­
and student jobs are related to the field of study. tution in the same field of study generates similar crowds. To test this
Conversely, we also observe pre-study work experience and student assumption, we estimate our results of wage with a nearest neighbour
jobs that are unrelated to the field of study. We conduct this analysis to matching method. The observables on which we match HE graduates are
show that our results for the other work experience sources are in line age, gender, type of education, level of education, subject and grades.
with the argument that only related work experience is helpful (e.g. Geel Our new results show that when we compare similar HE graduates with
& Backes-Gellner 2012; Baert et al. 2016). Drawing on Egg (2016), who different educational backgrounds, work experience from VET affects
relies on the experience classification framework of Quiñones et al. wage significantly and positively (estimated coefficient appear in
(1995), we hypothesise that the coefficient for VET should be about the Table A.13 in appendix A.3). Hence, our assumption about the gradu­
same size as student internships and related student jobs, smaller than ates’ comparability is reasonable.
work-based higher education, and larger than pre-study employment.
Table 2 shows the estimated coefficients for wage. For the tobit esti­
5. Conclusion
mation, work experience from VET has about the double the effect of
student internship and a slightly larger effect than student job related to
This study analyses the effect of VET work experience on the labour
the field of study. Work experience from VET is larger than that from pre-
market outcomes of HE graduates. Our findings show that having VET
study employment related to the field of study and lightly larger than
work experience increases wages and diminishes search time but does
work-based education. However, the results from the IV estimation are
not reduce the probability of internships one year after graduation.
puzzling. Differentiating between related and unrelated pre-study jobs for
Although this beneficial effect persists for five years after graduation, it
some reason influences the results for both VET and student internship,
is no longer significant for wage, unemployment spells or employment
although why the differentiation has such an influence remains unclear.
positions. The analysis of possible channels shows that screening, human
Therefore, we do not interpret those estimations. Nevertheless, from the
capital and signalling, not social network, are the prevalent channels.
first estimation, which shows slightly larger results but still comparable to
The benefits of work experience from VET one year after graduation
those in the literature, we argue that we are measuring work experience.
are sizeable. The estimated effect for wages ranges between 7% higher
wages for HE graduates with a VET diploma in the tobit model with
4.5.2. Are the results representative for higher education graduates?
controls and 19% higher wages in the IV model. Thus in Switzerland HE
In our analysis we focus on HE graduates with bachelor or master
graduates with VET earn between CHF 4800 (∼ 4470 Euro) and CHF
degrees who enter the labour market and do not continue their studies.19
13,100 (∼ 12,200 Euro) more one year after graduation than their
Selection might occur, in the sense that more able students might
counterparts with ACC. Having work experience from VET reduces
continue studying and might have ACC. If so, our findings for VET would
search time by two months in the IV model, indicating that HE graduates
be upwards biased. To look for any possible bias, we compare the
with VET search three months for a job and HE graduates with ACC need
summary statistics of the entire dataset with our sample (Table A.11 in
to search for five months. For internship, work experience from VET
appendix A.3 contains the results).
significantly reduces the probability of doing an internship after gradu­
Comparing the means of our dependent variables – wage, search time,
ation by 23.8% in the probit model but not significantly in the IV model.
internship – we see that the dataset has lower wages, higher search time
Comparing these coefficients to the coefficients of the other sources of
work experience makes the estimated coefficients from the IV model appear
somewhat large. However, given that we estimated the local average
19
As PhD graduates constitute a special case with too many unknowns, we treatment effect, the effects of the IV model are relevant only for HE grad­
therefore exclude them. uates who would have switched employers, not for all HE graduates,

11
M.E. Oswald-Egg and U. Renold
thereby likely explaining the large differences. Our findings are in line with
Shaw (2012) for VET in the UK. They are also in line with Nunley et al.
(2016) in the United States, Margaryan et al. (2019) in Germany and Bolli
et al. (2019a) in Switzerland for internships, San (1986) and Weiss et al.
(2014) for student jobs in Germany, and Böckerman et al. (2018) and
Böckerman et al. (2019) for vocational degrees in Finland identifying a
positive effect of work experience on labour market outcomes.
Five years after graduation, HE graduates with VET work experience
have higher wages by 2.7% in the tobit model (not significant) and 11.8%
in the IV model (significant) but are equally likely to be unemployed and
have no significant difference in their employment positions. This finding
of no negative effect for HE graduates with a VET diploma in the longer
run (five years after graduation) is in line with Hanushek et al. (2017),
who find no negative long-term effect of VET on labour market outcomes
after upper-secondary education in Switzerland. Moreover, we still find
positive effects that are simply not as robust as the results for one year
after graduation. However, the results contrast with the findings from
other countries of negative long-term effects of a VET diploma (Forster
et al., 2016; Hanushek et al., 2017; Hampf & Woessmann, 2017).
We have three explanations for these opposing results. First, our
sample focuses on highly educated individuals, and heterogeneity might
exist in the effect of VET, depending on the highest education level
attained. Second, as our time horizon is relatively short for the long run,
that negative effects appear only after prime age may be possible.
Nonetheless, we argue that the appearance of negative effects is unlikely
because all HE graduates have general skills due to their highest educa­
tional degree coming from higher education. Third, Switzerland’s VET
system is so well established that a VET diploma is not a disadvantage.
We find that HE graduates with VET have an advantage over their
academic peers at one year and similar outcomes after five years. We
have no reason to expect that HE graduates with ACC will surpass their
VET counterparts, because evidence shows that experience profiles
converge in the long term rather than invert (Mincer, 1974). In addition,
as all HE graduates have the same highest education level, we should not
expect any differences in wages due to education.
Overall, our results suggest that – for individuals who are not sure
whether to continue their education with a VET diploma or an academic
diploma – starting with a VET diploma is not a disadvantage, even when
continuing to higher education. Indeed, pursuing dual VET reduces the
foregone earnings during upper-secondary education, results in higher
wages after higher education, and helps HE graduates start earning
wages sooner. These findings extend the evidence in the literature that
labour market outcomes for VET are as good as or better than those for
academic education when one controls for years of education (Tuor &
Backes-Gellner, 2010) and demonstrates that VET is a good starting
point even for individuals planning to attend higher education. While
transitioning from higher education into the labour market can be a
major challenge, HE graduates with VET can use their work experience
to better capitalise on their higher education.
In line with the main findings in the literature (e.g. Weiss et al., 2014;
Nunley et al., 2016), we find the signalling channel to be important.
However, unlike the literature (except for Bolli et al., 2019a), we also find
the human capital channel to be important, joined to a lesser extent by
the screening channel. VET can plausibly be a signal, because entering
higher education requires some extra effort, which can constitute a signal
for higher productivity. That work experience from VET works through
the human capital channel is plausible insofar as VET conveys market­
able skills according to a structured curriculum developed in cooperation
with industry. Longer programme duration in VET – three to four years –
ensures the acquisition of a durable level of work experience. Moreover,
given that employers come to know the productivity level of HE gradu­
ates with VET, the screening channel is likewise plausible.
This study has two main policy implications. First, VET is an
appropriate and beneficial upper-secondary education even for students
who plan to later attend higher education. Policy-makers should
encourage VET in upper-secondary education not only for students who
12
Economics of Education Review 80 (2021) 102065
may not attend higher education but also for those who plan to do so. HE
graduates with VET benefit from increased wages and faster transitions
into the labour market, all without incurring foregone earnings during
upper-secondary education or lost time due to other methods of gaining
experience. This upper-secondary education pathway may indeed be an
important tool for improving equal opportunities, if VET helps young
people from a lower socio-economic background to reach upper-
secondary graduation and to afford higher education without having
to work on the side to gain work experience.
Second, certain VET programme design features are important for
driving positive outcomes. This second implication of our findings defines
the conditions under which such positive outcomes are possible. We find
that the VET programme structure has an important impact on later labour
market outcomes. Because the impact of VET is driven by human capital
acquisition, the skills developed during VET must be both relevant and of
high quality. These conditions imply that VET programmes need curricula
developed in cooperation with industry and that a substantial amount of
workplace learning is required. A school-based programme without any
exposure to work-based learning, a dual programme that is too short, or a
dual programme that is too narrow will not provide sufficient general
human capital to increase work experience and drive positive impacts.
A limitation of our identification strategy is that we cannot
completely rule out the self-selection of students into higher education.
Hence there is the possibility that HE graduates with VET, for example
have a higher ability in general and would therefore always have better
labour market outcomes. In such a case, our estimates would overstate
the benefit of VET. However, comparing HE graduates and including
various controls on personal characteristics, education, work experience
and employer characteristics largely minimise self-selection bias.
Another limitation of our results is that we are only able to causally
identify the local average treatment effect. Therefore, the external validity
for all HE graduates is not guaranteed, even though our results from the
explanatory analysis suggest that the effects are present but smaller. Our
results also only apply to HE graduates who enter the labour market and do
not continue their studies. In addition, the possibility of generalising our
results to other countries depends on the comparability of their education
systems to the Swiss system, especially in terms of the strong linkage be­
tween actors of the education and employment systems as well as the
substantial amount of workplace training VET graduates gain. Compara­
ble countries for education systems and the education-employment link­
age include Austria, Denmark and Germany. Finally, the advantages of
VET might arise only in countries with a permeable education system, in
which individuals with a VET diploma have the same access to educational
possibilities as individuals with an academic diploma. Despite these lim­
itations, our findings on VET work experience show that VET does not
need to be a second choice education pathway in an education system,
when the VET pathway is properly included in the system.
Declarations of Interest
Prof. Dr. Ursula Renold is chair of the university board of the Uni­
versity of Applied Sciences and Arts Northwestern Switzerland FHNW.
Maria Esther Oswald-Egg has none.
CRediT authorship contribution statement
Maria Esther Oswald-Egg: Conceptualization, Methodology,
Formal analysis, Writing - original draft, Writing - review & editing,
Visualization, Project administration. Ursula Renold: Conceptualiza­
tion, Writing - review & editing, Supervision, Funding acquisition.
Appendix A
A1. Descriptive statistics
 M.E. Oswald-Egg and U. Renold
Table A1
Correlation matrix 1 year post-graduation.
wage sear inta VET rVET Age fem faedu moedu mars child UAS mast inst subj sem grad wave prewe ints stuwe wbe ex publ size noga empr rlVET net screen

wage 1.00
sear -0.17 1.00
inta -0.26 0.11 1.00
VET 0.20 -0.07 -0.16 1.00
rVET 0.17 -0.05 -0.04 0.23 1.00
Age 0.13 -0.04 -0.06 0.08 0.15 1.00
fem -0.23 0.05 0.08 -0.25 -0.06 -0.16 1.00
faedu 0.02 -0.01 0.03 -0.08 0.07 0.01 -0.02 1.00
moedu -0.03 -0.01 0.04 -0.12 -0.03 0.00 0.01 0.47 1.00
mars 0.01 0.00 -0.06 0.02 0.01 0.10 0.01 -0.02 -0.03 1.00
child -0.02 0.00 -0.04 0.03 0.01 0.11 -0.04 -0.01 -0.02 0.33 1.00
UAS 0.03 -0.05 -0.20 0.54 0.12 -0.09 -0.03 -0.12 -0.12 0.04 0.04 1.00
mast 0.07 0.01 0.10 -0.43 -0.11 0.26 -0.02 0.12 0.11 -0.02 -0.02 -0.76 1.00
inst 0.11 -0.08 -0.21 0.49 0.23 -0.04 -0.09 -0.06 -0.08 0.02 0.03 0.83 -0.59 1.00
subj 0.22 -0.08 -0.19 0.29 0.07 -0.03 -0.32 -0.00 -0.02 -0.01 -0.00 0.21 -0.10 0.31 1.00
sem -0.04 0.02 -0.09 0.14 0.05 0.02 0.03 -0.05 -0.04 0.04 0.04 0.28 -0.45 0.20 -0.02 1.00
13

grad 0.01 -0.02 -0.00 -0.18 -0.06 0.00 0.08 0.05 0.07 0.01 -0.01 -0.21 0.30 -0.19 -0.03 -0.17 1.00
wave -0.04 -0.00 0.00 -0.03 -0.07 0.27 0.02 -0.00 0.05 0.03 0.04 0.01 0.04 0.00 -0.02 0.03 0.05 1.00
prewe 0.03 -0.03 -0.02 0.09 0.09 0.13 0.00 0.01 -0.01 0.02 0.02 0.11 -0.11 0.09 -0.05 0.05 -0.03 0.02 1.00
ints -0.09 0.01 -0.05 -0.32 -0.13 -0.05 0.25 0.03 0.07 0.02 0.00 -0.18 0.16 -0.16 -0.07 -0.03 0.15 0.00 -0.04 1.00
stuwe 0.01 0.01 0.02 0.02 0.07 0.06 0.08 0.01 -0.00 0.01 0.01 0.01 -0.01 -0.02 -0.13 0.07 0.00 -0.00 0.22 -0.01 1.00
wbe 0.17 -0.05 -0.13 0.25 0.12 0.22 -0.05 -0.05 -0.06 0.07 0.06 0.27 -0.16 0.25 -0.04 0.26 -0.07 0.08 0.10 -0.25 0.14 1.00
ex 0.01 0.00 0.08 -0.07 -0.02 0.01 0.02 0.03 0.04 -0.03 -0.01 -0.09 0.09 -0.05 -0.10 -0.03 0.07 -0.01 0.01 -0.01 0.03 -0.06 1.00
publ -0.17 0.05 0.00 -0.10 -0.06 -0.04 0.23 -0.06 -0.03 0.03 0.02 0.03 -0.02 -0.05 -0.11 0.02 0.10 0.04 -0.02 0.18 0.00 -0.06 -0.06 1.00
size 0.22 -0.03 0.00 0.02 0.04 0.01 -0.04 0.01 -0.00 -0.01 -0.00 0.01 0.01 0.02 0.03 -0.03 0.01 0.02 0.00 0.01 0.01 0.05 0.03 0.12 1.00
noga -0.25 0.05 0.03 -0.21 -0.11 -0.04 0.29 -0.03 0.02 0.03 0.01 -0.03 0.00 -0.09 -0.27 0.04 0.11 0.05 0.00 0.23 0.02 -0.05 -0.01 0.37 -0.15 1.00
empr 0.06 -0.02 -0.00 0.12 0.47 0.07 -0.04 0.04 -0.01 -0.02 -0.02 0.05 -0.06 0.23 0.04 0.00 -0.06 -0.03 0.02 -0.10 -0.02 0.05 -0.00 -0.08 0.01 -0.10 1.00
rlVET 0.20 -0.07 -0.15 0.78 0.17 -0.01 -0.23 -0.07 -0.10 0.01 0.02 0.43 -0.33 0.38 0.28 0.10 -0.13 -0.04 0.07 -0.32 0.02 0.24 -0.05 -0.11 0.03 -0.20 0.09 1.00
net -0.03 0.03 -0.04 -0.04 -0.03 0.01 0.02 0.02 0.02 -0.00 0.01 -0.02 0.01 -0.02 -0.02 0.01 -0.00 -0.01 0.01 0.06 0.02 -0.01 -0.01 -0.01 -0.05 0.05 -0.01 -0.04 1.00
screen 0.09 -0.31 -0.12 0.11 0.02 0.09 -0.08 -0.03 -0.02 0.03 0.03 0.10 -0.08 0.10 0.03 0.11 -0.02 0.04 0.05 -0.08 0.04 0.30 -0.04 -0.02 -0.03 -0.02 0.02 0.11 0.00 1.00

Economics of Education Review 80 (2021) 102065

Note: The table shows the correlation matrix for the data of the one-year post-graduation surveys. The abbreviations stand for the variables wage (wage), search time (sear), internship after graduation (inta), vocational
education and training diploma (VET), regional enrolment rate in VET (rVET), female (fem), education of father (faedu), education of mother (moedu), marital status (mars), child (child), type of education being
university of applied science (UAS), education level being master level (mast), institution (inst), subject (subj), number of semesters (sem), grade (grad), cohort (wave), work experience gained before higher education
(prewe), student internship (ints), student job (stuwe), workbased education (wbe), exchange during studies (ex), public employment (publ), size (size), noga1 (noga), region of employer (empr), VET related to studies
and/or employer (rlVET), employment found through social network (net), employment found at previous employer (screen).
 M.E. Oswald-Egg and U. Renold
Table A2
Correlation matrix 5 year post-graduation.
wage unemp pos VET rVET Age fem faedu moedu mars child UAS mast inst subj wave constud addstud conedu publ size noga empr

wage 1.00
unemp -0.10 1.00
pos 0.26 -0.10 1.00
VET 0.10 -0.07 0.06 1.00
rVET 0.10 -0.21 0.09 0.21 1.00
Age 0.15 0.04 0.08 0.01 0.14 1.00
fem -0.29 0.03 -0.09 -0.30 -0.02 -0.10 1.00
faedu 0.05 -0.06 0.03 -0.06 0.07 0.03 -0.01 1.00
moedu -0.01 -0.02 0.01 -0.12 -0.04 -0.02 0.06 0.44 1.00
mars -0.06 -0.03 0.02 0.02 -0.01 0.12 0.01 -0.01 -0.01 1.00
child -0.11 -0.03 -0.00 0.02 -0.03 0.14 0.01 -0.03 -0.02 0.56 1.00
UAS -0.10 -0.08 -0.00 0.58 0.16 -0.11 -0.06 -0.10 -0.09 0.02 0.05 1.00
14

mast 0.18 0.07 0.06 -0.46 -0.14 0.25 0.01 0.10 0.08 0.01 -0.02 -0.79 1.00
inst 0.00 -0.11 0.02 0.51 0.24 -0.06 -0.12 -0.04 -0.06 0.02 0.03 0.83 -0.62 1.00
subj 0.13 -0.13 0.02 0.32 0.06 -0.04 -0.38 0.02 -0.04 0.01 0.01 0.26 -0.11 0.34 1.00
wave -0.02 0.02 -0.02 -0.09 0.01 0.28 0.00 0.04 0.04 0.06 0.06 -0.01 0.04 -0.01 -0.02 1.00
constud -0.12 -0.02 -0.05 0.02 0.04 -0.10 -0.02 0.02 0.01 -0.08 -0.07 0.00 -0.13 0.00 -0.05 -0.00 1.00
addstud 0.01 -0.01 0.01 -0.04 -0.03 -0.01 0.07 0.01 0.01 -0.01 -0.02 -0.03 0.03 -0.05 -0.05 -0.01 -0.09 1.00
conedu 0.01 -0.05 0.06 0.00 0.07 0.01 0.06 0.04 0.01 -0.01 -0.02 0.06 -0.05 0.04 -0.07 0.02 -0.11 -0.06 1.00
publ -0.22 -0.02 -0.13 -0.14 -0.06 -0.06 0.26 -0.03 0.01 0.05 0.07 0.04 -0.07 -0.04 -0.13 0.01 -0.00 0.08 0.09 1.00
size 0.19 -0.03 -0.16 0.00 0.03 -0.00 -0.03 0.02 0.01 0.00 -0.00 0.00 0.01 0.01 0.05 -0.01 0.03 0.04 0.01 0.13 1.00
noga -0.27 0.00 -0.09 -0.25 -0.12 -0.06 0.33 -0.03 0.02 0.04 0.05 -0.04 -0.01 -0.11 -0.32 -0.01 0.05 0.10 0.07 0.38 -0.17 1.00
empr 0.07 -0.12 0.08 0.14 0.51 0.07 -0.02 0.03 -0.04 -0.03 -0.06 0.08 -0.10 0.24 0.04 0.00 0.04 -0.05 0.03 -0.10 0.01 -0.12 1.00

Note: The table shows the correlation matrix for the data of the five-year post-graduation surveys. The abbreviations stand for the variables wage (wage), unemployment (unemp), position (pos), vocational education and
training diploma (VET), regional enrolment rate in VET (rVET), female (fem), education of father (faedu), education of mother (moedu), marital status (mars), child (child), type of education being university of applied
science (UAS), education level being master level (mast), institution (inst), subject (subj), wave (wave), continuing studies (constud), additional studies (conadd), continuing education (conedu), work experience gained

Economics of Education Review 80 (2021) 102065

before higher education (prewe), student internship (ints), student job (stuwe), workbased education (wbe), exchange during studies (ex), public employment (publ), size (size), noga1 (noga), region of employer (empr).
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A3 Table A3 (continued )

Summary statistics 1 year post-graduation. ACC VET
ACC VET
Mean SD Obs Mean SD Obs
Mean SD Obs Mean SD Obs
CV: Experience
Dependent variables Pre-study job 0.86 0.35 8558 0.91 0.28 5341
Wage in KCHF 69.03 25.07 8558 78.70 20.67 5341 Student internship 0.59 0.49 8558 0.26 0.44 5341
Search time 4.74 5.41 8558 3.98 4.80 5341 Student job 0.87 0.34 8558 0.88 0.32 5341
Internship 0.22 0.41 8558 0.09 0.29 5341 Work-based high. edu. 0.06 0.24 8558 0.23 0.42 5341
Instrumental variable Exchange 0.18 0.38 8558 0.13 0.33 5341
Regional enrolment rate in 0.70 0.12 8558 0.75 0.09 5341 CV: Employer
VET characteristics
CV: Personal Public empl. 0.36 0.48 8558 0.27 0.44 5341
characteristics Company size
Age 26.10 1.51 8558 26.35 1.73 5341 Small 0.34 0.47 8558 0.32 0.47 5341
Female 0.60 0.49 8558 0.34 0.47 5341 Medium 0.20 0.40 8558 0.20 0.40 5341
Education of father Large 0.46 0.50 8558 0.48 0.50 5341
Dad: Comp. 0.08 0.26 8558 0.08 0.27 5341 NOGA1
Dad: Upp.-sec. 0.38 0.49 8558 0.47 0.50 5341 Agriculture ... 0.00 0.01 8558 0.00 0.04 5341
Dad: Post-sec. 0.54 0.50 8558 0.45 0.50 5341 Mining ... 0.00 0.00 8558 0.00 0.02 5341
Education of mother Manufacturing 0.09 0.28 8558 0.20 0.40 5341
Mom: Comp. 0.09 0.28 8558 0.10 0.30 5341 Electricity ... 0.01 0.08 8558 0.01 0.10 5341
Mom: Upp.-sec. 0.56 0.50 8558 0.68 0.47 5341 Water Supply ... 0.00 0.04 8558 0.00 0.04 5341
Mom: Post-sec. 0.35 0.48 8558 0.22 0.41 5341 Construction 0.01 0.09 8558 0.02 0.13 5341
Marital status 0.06 0.24 8558 0.07 0.26 5341 Wholesale ... 0.07 0.26 8558 0.06 0.23 5341
Child 0.02 0.13 8558 0.03 0.16 5341 Transportation ... 0.02 0.15 8558 0.03 0.16 5341
CV: Education Accommodation ... 0.01 0.12 8558 0.01 0.08 5341
UAS 0.41 0.49 8558 0.96 0.21 5341 Information ... 0.07 0.26 8558 0.10 0.30 5341
Master 0.51 0.50 8558 0.09 0.28 5341 Financial ... 0.08 0.27 8558 0.10 0.29 5341
Institute Real Estate 0.00 0.07 8558 0.01 0.08 5341
UNI 1 0.04 0.20 8558 0.00 0.05 5341 Scientific ... 0.22 0.42 8558 0.24 0.43 5341
UNI 2 0.06 0.24 8558 0.01 0.08 5341 Administrative ... 0.03 0.16 8558 0.02 0.15 5341
UNI 3 0.06 0.23 8558 0.00 0.06 5341 Public ... 0.06 0.23 8558 0.03 0.17 5341
UNI 4 0.04 0.20 8558 0.00 0.05 5341 Education 0.08 0.28 8558 0.05 0.22 5341
UNI 5 0.05 0.23 8558 0.00 0.06 5341 Health ... 0.18 0.38 8558 0.11 0.31 5341
UNI 6 0.01 0.08 8558 0.00 0.02 5341 Arts ... 0.03 0.17 8558 0.01 0.09 5341
UNI 7 0.02 0.15 8558 0.00 0.05 5341 Other service ... 0.03 0.17 8558 0.01 0.11 5341
UNI 8 0.04 0.19 8558 0.00 0.07 5341 Extraterritorial ... 0.01 0.07 8558 0.00 0.01 5341
UNI 9 0.08 0.27 8558 0.01 0.08 5341 Employer region
UNI 10 0.01 0.12 8558 0.00 0.02 5341 Geneva 0.08 0.27 8558 0.02 0.15 5341
UNI 11 0.05 0.23 8558 0.00 0.07 5341 Lausanne 0.12 0.32 8558 0.06 0.23 5341
UNI 12 0.12 0.32 8558 0.01 0.08 5341 Sion 0.03 0.17 8558 0.02 0.15 5341
UAS 1 0.05 0.21 8558 0.12 0.32 5341 Fribourg 0.03 0.18 8558 0.03 0.16 5341
UAS 2 0.16 0.37 8558 0.17 0.38 5341 Neuchatel 0.03 0.16 8558 0.02 0.14 5341
UAS 3 0.04 0.20 8558 0.13 0.34 5341 Biel 0.03 0.16 8558 0.04 0.19 5341
UAS 4 0.03 0.18 8558 0.11 0.32 5341 Bern 0.11 0.32 8558 0.12 0.33 5341
UAS 5 0.02 0.13 8558 0.02 0.15 5341 Basel 0.06 0.24 8558 0.05 0.22 5341
UAS 6 0.03 0.18 8558 0.13 0.34 5341 Aarau-Olten 0.03 0.17 8558 0.05 0.21 5341
UAS 7 0.08 0.27 8558 0.26 0.44 5341 Zurich 0.29 0.45 8558 0.34 0.47 5341
UAS 8 0.00 0.02 8558 0.01 0.08 5341 Winterthur-SH 0.02 0.15 8558 0.06 0.24 5341
Subject St. Gallen 0.03 0.17 8558 0.07 0.26 5341
Humanities 0.05 0.21 8558 0.01 0.08 5341 Chur 0.01 0.12 8558 0.02 0.14 5341
Arts 0.04 0.20 8558 0.03 0.16 5341 Luzern 0.04 0.20 8558 0.06 0.25 5341
Edu. Sciences 0.20 0.40 8558 0.06 0.23 5341 Bellinzona 0.01 0.12 8558 0.01 0.09 5341
Economics 0.27 0.44 8558 0.33 0.47 5341 Lugano 0.02 0.15 8558 0.01 0.10 5341
Nat. Sciences 0.10 0.30 8558 0.01 0.08 5341 Abroad 0.04 0.19 8558 0.02 0.14 5341
Medi. Sciences 0.03 0.17 8558 0.00 0.04 5341 Channel
Health 0.10 0.30 8558 0.07 0.26 5341 Related VET 0.00 0.00 8558 0.71 0.45 5341
Engineering 0.08 0.27 8558 0.11 0.31 5341 Network 0.22 0.41 7461 0.19 0.39 4258
Agri. Sciences 0.11 0.31 8558 0.39 0.49 5341 Screening 0.08 0.28 8476 0.15 0.36 5295
Others 0.02 0.13 8558 0.00 0.05 5341
Note: The table shows the unweighted summary statistics including mean,
No. of semester 5.97 2.17 8558 6.52 1.44 5341
Grade 5.14 0.39 8558 5.00 0.38 5341 standard deviation (SD), and number of observations (Obs) for the data of the
Cohort one-year post-graduation surveys. The columns under ACC display the statistics
2011 0.10 0.29 8558 0.12 0.32 5341 for higher education graduates with an academic diploma in upper-secondary
2013 0.24 0.43 8558 0.22 0.41 5341 education. The columns under VET display the statistics for higher education
2015 0.28 0.45 8558 0.32 0.47 5341 graduates with a VET diploma in upper-secondary education.
2017 0.38 0.49 8558 0.35 0.48 5341

15
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A4 Table A4 (continued )

Summary statistics 5 years post-graduation. ACC VET
ACC VET
Mean SD Obs Mean SD Obs
Mean SD Obs Mean SD Obs
2015 0.31 0.46 1991 0.39 0.49 1327
Wage in KCHF 83.90 31.97 1991 90.43 29.73 1327 2017 0.69 0.46 1991 0.61 0.49 1327
Unemployment 0.20 0.40 1991 0.14 0.35 1327 Cont. studies 0.12 0.32 1991 0.13 0.34 1327
Position 2.49 0.77 1991 2.60 0.81 1327 Add. studies 0.11 0.32 1991 0.09 0.29 1327
Instrumental variable Cont. education 0.12 0.33 1991 0.12 0.33 1327
Regional enrolment rate in 0.71 0.11 1991 0.75 0.09 1327 CV: Employer
VET characteristics
CV: Personal Public empl. 0.35 0.48 1991 0.22 0.41 1327
characteristics Company size
Age 29.67 1.23 1991 29.69 1.31 1327 Small 0.32 0.47 1991 0.30 0.46 1327
Female 0.62 0.48 1991 0.31 0.46 1327 Medium 0.19 0.39 1991 0.21 0.41 1327
Education of father Large 0.50 0.50 1991 0.49 0.50 1327
Dad: Comp. 0.07 0.25 1991 0.07 0.26 1327 NOGA1
Dad: Upp.-sec. 0.40 0.49 1991 0.47 0.50 1327 Manufacturing 0.10 0.30 1991 0.25 0.43 1327
Dad: Post-sec. 0.53 0.50 1991 0.46 0.50 1327 Electricity ... 0.01 0.09 1991 0.01 0.12 1327
Education of mother Water Supply ... 0.00 0.05 1991 0.00 0.00 1327
Mom: Comp. 0.09 0.29 1991 0.11 0.31 1327 Construction 0.01 0.09 1991 0.01 0.12 1327
Mom: Upp.-sec. 0.60 0.49 1991 0.70 0.46 1327 Wholesale ... 0.07 0.25 1991 0.08 0.26 1327
Mom: Post-sec. 0.31 0.46 1991 0.19 0.40 1327 Transportation ... 0.03 0.16 1991 0.02 0.16 1327
Marital status 0.20 0.40 1991 0.22 0.41 1327 Accommodation ... 0.01 0.08 1991 0.00 0.07 1327
Child 0.12 0.33 1991 0.14 0.34 1327 Information ... 0.08 0.27 1991 0.11 0.32 1327
CV: Education Financial ... 0.08 0.27 1991 0.09 0.28 1327
UAS 0.40 0.49 1991 0.97 0.17 1327 Real Estate 0.00 0.06 1991 0.01 0.08 1327
Master 0.49 0.50 1991 0.05 0.22 1327 Scientific ... 0.21 0.41 1991 0.21 0.41 1327
Institute Administrative ... 0.02 0.14 1991 0.02 0.12 1327
UNI 1 0.05 0.21 1991 0.00 0.03 1327 Public ... 0.07 0.26 1991 0.04 0.19 1327
UNI 2 0.07 0.25 1991 0.00 0.06 1327 Education 0.05 0.21 1991 0.03 0.17 1327
UNI 3 0.05 0.21 1991 0.00 0.05 1327 Health ... 0.21 0.40 1991 0.09 0.28 1327
UNI 4 0.05 0.22 1991 0.00 0.05 1327 Arts ... 0.02 0.15 1991 0.01 0.10 1327
UNI 5 0.06 0.24 1991 0.01 0.07 1327 Other service ... 0.04 0.19 1991 0.02 0.13 1327
UNI 6 0.01 0.08 1991 0.00 0.00 1327 Extraterritorial ... 0.00 0.06 1991 0.00 0.04 1327
UNI 7 0.03 0.18 1991 0.00 0.07 1327 Employer region
UNI 8 0.03 0.16 1991 0.00 0.03 1327 Geneva 0.07 0.26 1991 0.03 0.16 1327
UNI 9 0.07 0.26 1991 0.00 0.07 1327 Lausanne 0.13 0.34 1991 0.06 0.24 1327
UNI 10 0.01 0.12 1991 0.00 0.00 1327 Sion 0.04 0.19 1991 0.03 0.17 1327
UNI 11 0.06 0.24 1991 0.00 0.05 1327 Fribourg 0.03 0.17 1991 0.03 0.16 1327
UNI 12 0.12 0.32 1991 0.00 0.04 1327 Neuchatel 0.02 0.15 1991 0.03 0.18 1327
UAS 1 0.05 0.22 1991 0.12 0.33 1327 Biel 0.03 0.18 1991 0.05 0.21 1327
UAS 2 0.16 0.36 1991 0.19 0.39 1327 Bern 0.12 0.33 1991 0.11 0.31 1327
UAS 3 0.03 0.17 1991 0.13 0.34 1327 Basel 0.06 0.23 1991 0.04 0.21 1327
UAS 4 0.03 0.18 1991 0.11 0.31 1327 Aarau-Olten 0.03 0.17 1991 0.05 0.22 1327
UAS 5 0.01 0.12 1991 0.03 0.17 1327 Zurich 0.29 0.46 1991 0.32 0.47 1327
UAS 6 0.03 0.17 1991 0.15 0.36 1327 Winterthur-SH 0.03 0.17 1991 0.06 0.24 1327
UAS 7 0.09 0.28 1991 0.23 0.42 1327 St. Gallen 0.04 0.19 1991 0.06 0.24 1327
UAS 8 0.00 0.00 1991 0.00 0.06 1327 Chur 0.02 0.12 1991 0.02 0.15 1327
Subject Luzern 0.05 0.22 1991 0.06 0.24 1327
Humanities 0.05 0.21 1991 0.01 0.09 1327 Bellinzona 0.01 0.12 1991 0.01 0.10 1327
Arts 0.03 0.17 1991 0.02 0.13 1327 Lugano 0.02 0.15 1991 0.02 0.15 1327
Edu. Sciences 0.23 0.42 1991 0.05 0.21 1327 Abroad 0.01 0.07 1991 0.01 0.10 1327
Economics 0.24 0.43 1991 0.31 0.46 1327
Note: The table shows the unweighted summary statistics including mean,
Nat. Sciences 0.09 0.29 1991 0.00 0.07 1327
Medi. Sciences 0.03 0.18 1991 0.00 0.00 1327 standard deviation (SD), and number of observations (Obs) for the data of the
Health 0.11 0.32 1991 0.06 0.24 1327 five-year post-graduation surveys. The columns under ACC display the statistics
Engineering 0.07 0.26 1991 0.09 0.29 1327 for higher education graduates with an academic diploma in upper-secondary
Agri. Sciences 0.13 0.33 1991 0.45 0.50 1327 education. The columns under VET display the statistics for higher education
Others 0.02 0.14 1991 0.00 0.03 1327 graduates with a VET diploma in upper-secondary education.
Cohort

16
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

A2. Complete estimation results

Table A5
Regression table for wage one year post-graduation.
Tobit1 Tobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 3.222*** (0.11) 1.354*** (0.33) 3.654*** (0.71) 0.901 (1.01)
B. Second stage
VET 0.169*** (0.01) 0.069*** (0.01) 0.407*** (0.01) 0.189*** (0.03) 0.289*** (0.01) 0.192*** (0.03)
Age 0.013*** (0.00) 0.009*** (0.00) 0.008*** (0.00)
Female -0.052*** (0.01) -0.041*** (0.01) -0.040*** (0.01)
Dad: Upp.-sec. -0.000 (0.01) -0.004 (0.01) -0.004 (0.01)
Dad: Post-sec. 0.001 (0.01) -0.001 (0.01) -0.001 (0.01)
Mom: Upp.-sec. -0.004 (0.01) -0.005 (0.01) -0.005 (0.01)
Mom: Post-sec. -0.012 (0.01) -0.007 (0.01) -0.007 (0.01)
Marital status 0.020* (0.01) 0.020* (0.01) 0.021* (0.01)
Child -0.077*** (0.02) -0.076*** (0.02) -0.076*** (0.02)
Res: BE -0.053*** (0.02) -0.052*** (0.02) -0.058*** (0.02)
Res: LU -0.018 (0.02) -0.013 (0.02) -0.029 (0.02)
Res: UR -0.027 (0.05) -0.024 (0.05) -0.015 (0.08)
Res: SZ -0.009 (0.02) -0.002 (0.02) 0.003 (0.03)
Res: OW -0.129* (0.07) -0.122* (0.07) -0.134* (0.07)
Res: NW -0.141** (0.06) -0.132** (0.06) -0.082 (0.07)
Res: GL -0.035 (0.05) -0.022 (0.05) 0.045 (0.07)
Res: ZG 0.019 (0.02) 0.018 (0.02) 0.025 (0.03)
Res: FR -0.042* (0.02) -0.043* (0.03) -0.077*** (0.03)
Res: SO 0.014 (0.02) 0.014 (0.02) -0.012 (0.02)
Res: BS -0.067** (0.03) -0.058** (0.03) -0.086*** (0.03)
Res: BL 0.003 (0.03) 0.011 (0.03) -0.045 (0.03)
Res: SH -0.047 (0.03) -0.050 (0.03) -0.026 (0.04)
Res: AR -0.034 (0.05) -0.041 (0.05) -0.047 (0.06)
Res: AI 0.033 (0.11) 0.031 (0.10) -0.050 (0.16)
Res: SG 0.011 (0.02) 0.006 (0.02) 0.006 (0.02)
Res: GR -0.046 (0.03) -0.041 (0.03) -0.028 (0.03)
Res: AG -0.003 (0.01) -0.008 (0.01) -0.023 (0.02)
Res: TG 0.001 (0.02) -0.004 (0.02) -0.015 (0.02)
Res: TI -0.092*** (0.03) -0.087*** (0.03) -0.052 (0.05)
Res: VD -0.080*** (0.03) -0.078*** (0.03) -0.096*** (0.03)
Res: VS -0.013 (0.03) -0.006 (0.03) -0.029 (0.03)
Res: NE -0.107*** (0.04) -0.103*** (0.04) -0.101** (0.04)
Res: GE -0.135*** (0.03) -0.123*** (0.04) -0.166*** (0.04)
Res: JU -0.067* (0.03) -0.069** (0.03) -0.079* (0.05)
UAS 0.228*** (0.04) 0.165*** (0.05) 0.168*** (0.05)
Master 0.158*** (0.02) 0.173*** (0.02) 0.174*** (0.02)
UNI 2 0.006 (0.03) 0.004 (0.03) 0.007 (0.03)
UNI 3 0.101*** (0.03) 0.097*** (0.03) 0.097*** (0.03)
UNI 4 0.043 (0.04) 0.031 (0.04) 0.028 (0.04)
UNI 5 -0.080** (0.04) -0.084** (0.04) -0.084** (0.04)
UNI 6 0.025 (0.07) 0.021 (0.07) 0.026 (0.07)
UNI 7 -0.065 (0.04) -0.076* (0.04) -0.076* (0.04)
UNI 8 0.189*** (0.03) 0.190*** (0.03) 0.193*** (0.03)
UNI 9 0.029 (0.03) 0.030 (0.03) 0.036 (0.03)
UNI 10 -0.064 (0.05) -0.063 (0.05) -0.057 (0.05)
UNI 11 0.032 (0.03) 0.041 (0.03) 0.038 (0.03)
UNI 12 0.077*** (0.03) 0.093*** (0.03) 0.096*** (0.03)
UAS 1 -0.075** (0.03) -0.063* (0.03) -0.065* (0.03)
UAS 2 -0.079** (0.03) -0.051 (0.04) -0.056 (0.04)
UAS 3 -0.076** (0.03) -0.065* (0.04) -0.070** (0.04)
UAS 4 -0.069** (0.03) -0.063* (0.04) -0.066* (0.04)
UAS 5 -0.127*** (0.04) -0.109** (0.05) -0.107** (0.05)
UAS 6 -0.100*** (0.03) -0.089** (0.04) -0.086** (0.04)
UAS 7 -0.089*** (0.03) -0.081** (0.03) -0.079** (0.03)
UAS 8 0.000 (.)
Arts -0.333*** (0.04) -0.317*** (0.04) -0.317*** (0.04)
Edu. Sciences 0.150*** (0.03) 0.142*** (0.03) 0.142*** (0.03)
Economics 0.229*** (0.03) 0.217*** (0.03) 0.215*** (0.03)
Nat. Sciences 0.147*** (0.03) 0.132*** (0.03) 0.131*** (0.03)
Medi. Sciences 0.238*** (0.04) 0.216*** (0.04) 0.216*** (0.04)
Health 0.240*** (0.03) 0.234*** (0.03) 0.231*** (0.03)
Engineering 0.209*** (0.03) 0.180*** (0.03) 0.179*** (0.03)
Agri. Sciences 0.245*** (0.03) 0.214*** (0.03) 0.213*** (0.03)
Others 0.028 (0.05) 0.011 (0.05) 0.013 (0.05)
No. of semester 0.000 (0.00) 0.002 (0.00) 0.002 (0.00)
Grade 0.054*** (0.01) 0.051*** (0.01) 0.050*** (0.01)
Cohort=2013 -0.062*** (0.01) -0.052*** (0.01) -0.052*** (0.01)
Cohort=2015 -0.071*** (0.01) -0.063*** (0.01) -0.063*** (0.01)
(continued on next page)
17
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A5 (continued )
Tobit1 Tobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Cohort=2017 -0.080*** (0.01) -0.070*** (0.01) -0.070*** (0.01)

Pre-study job 0.004 (0.01) 0.006 (0.01) 0.007 (0.01)
Student internship 0.034*** (0.01) 0.047*** (0.01) 0.047*** (0.01)
Student job -0.006 (0.01) -0.008 (0.01) -0.009 (0.01)
Work-based high. edu. 0.066*** (0.01) 0.058*** (0.01) 0.059*** (0.01)
Exchange -0.003 (0.01) -0.003 (0.01) -0.003 (0.01)
Public empl. -0.015 (0.01) -0.016* (0.01) -0.016* (0.01)
Medium 0.081*** (0.01) 0.081*** (0.01) 0.081*** (0.01)
Large 0.121*** (0.01) 0.121*** (0.01) 0.120*** (0.01)
Manufacturing 0.317** (0.16) 0.338** (0.16) 0.338** (0.15)
Electricity ... 0.405** (0.16) 0.424*** (0.16) 0.426*** (0.16)
Water Supply ... 0.411** (0.16) 0.436*** (0.16) 0.433*** (0.16)
Construction 0.323** (0.16) 0.339** (0.16) 0.338** (0.16)
Wholesale ... 0.298* (0.16) 0.323** (0.16) 0.325** (0.16)
Transportation ... 0.240 (0.16) 0.273* (0.16) 0.274* (0.16)
Accommodation ... -0.016 (0.17) 0.021 (0.17) 0.023 (0.16)
Information ... 0.318** (0.16) 0.341** (0.16) 0.342** (0.16)
Financial ... 0.390** (0.16) 0.414*** (0.16) 0.415*** (0.16)
Real Estate 0.269 (0.17) 0.291* (0.16) 0.291* (0.16)
Scientific ... 0.321** (0.16) 0.346** (0.16) 0.347** (0.16)
Administrative ... 0.240 (0.16) 0.270* (0.16) 0.272* (0.16)
Public ... 0.234 (0.16) 0.264* (0.16) 0.264* (0.16)
Education 0.154 (0.16) 0.183 (0.16) 0.183 (0.16)
Health ... 0.204 (0.16) 0.236 (0.16) 0.238 (0.16)
Arts ... 0.087 (0.17) 0.120 (0.16) 0.120 (0.16)
Other service ... 0.088 (0.16) 0.118 (0.16) 0.120 (0.16)
Extraterritorial ... 0.088 (0.20) 0.120 (0.20) 0.128 (0.20)
Lausanne -0.090*** (0.02) -0.089*** (0.02) -0.087*** (0.02)
Sion -0.096*** (0.03) -0.102*** (0.03) -0.101*** (0.03)
Fribourg -0.088*** (0.03) -0.088*** (0.03) -0.089*** (0.03)
Neuchatel -0.074** (0.04) -0.079** (0.04) -0.077** (0.04)
Biel -0.038 (0.03) -0.043 (0.03) -0.043 (0.03)
Bern -0.047 (0.03) -0.052* (0.03) -0.052* (0.03)
Basel -0.096*** (0.03) -0.094*** (0.03) -0.093*** (0.03)
Aarau-Olten -0.068** (0.03) -0.066** (0.03) -0.067** (0.03)
Zurich -0.036 (0.03) -0.036 (0.03) -0.036 (0.03)
Winterthur-SH -0.020 (0.03) -0.025 (0.03) -0.027 (0.03)
St. Gallen -0.043 (0.04) -0.045 (0.04) -0.046 (0.04)
Chur -0.027 (0.04) -0.026 (0.04) -0.029 (0.04)
Luzern -0.036 (0.03) -0.036 (0.03) -0.036 (0.03)
Bellinzona -0.036 (0.05) -0.038 (0.05) -0.040 (0.05)
Lugano -0.184*** (0.05) -0.181*** (0.05) -0.183*** (0.05)
Abroad -0.249*** (0.06) -0.249*** (0.06) -0.248*** (0.06)
Ori: BE -0.061*** (0.01) 0.012 (0.02)
Ori: LU 0.030** (0.02) 0.028 (0.02)
Ori: UR 0.022 (0.04) -0.003 (0.06)
Ori: SZ 0.049** (0.02) 0.001 (0.03)
Ori: OW -0.052 (0.06) 0.020 (0.05)
Ori: NW -0.120* (0.06) -0.052 (0.06)
Ori: GL -0.056 (0.05) -0.075 (0.06)
Ori: ZG 0.048 (0.03) -0.009 (0.03)
Ori: FR -0.044** (0.02) 0.051** (0.02)
Ori: SO 0.003 (0.02) 0.040* (0.02)
Ori: BS -0.127*** (0.04) 0.020 (0.04)
Ori: BL 0.010 (0.02) 0.075*** (0.03)
Ori: SH -0.035 (0.03) -0.023 (0.03)
Ori: AR -0.030 (0.04) 0.012 (0.03)
Ori: AI 0.189*** (0.07) 0.110 (0.10)
Ori: SG 0.017 (0.02) 0.005 (0.02)
Ori: GR -0.041* (0.02) -0.010 (0.02)
Ori: AG 0.006 (0.01) 0.025 (0.02)
Ori: TG 0.004 (0.02) 0.017 (0.02)
Ori: TI -0.157*** (0.02) -0.028 (0.04)
Ori: VD -0.092*** (0.02) 0.027 (0.02)
Ori: VS -0.032** (0.02) 0.032 (0.03)
Ori: NE -0.151*** (0.03) 0.003 (0.03)
Ori: GE -0.034* (0.02) 0.059* (0.03)
Ori: JU -0.113*** (0.03) 0.019 (0.04)
Constant 11.069*** (0.01) 9.980*** (0.19) 10.977*** (0.01) 10.030*** (0.18) 11.056*** (0.01) 10.036*** (0.18)
Observations 13,571 13,571 13,571 13,571 13,571 13,571
Pseudo R2 0.038 0.336

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the tobit regressions (first two columns) and IV regressions (last four columns) on wage
(ln). The data is from the one-year post-graduation surveys. Columns Tobit1 and IV1 include no controls, whereas IV3 includes region of origin fixed effects. Columns Tobit2
and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first stage include the same controls as the estimations
from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance at the 1%, 5% and 10% level, respectively.

18
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A6
Regression table for search time of first post-graduation employment.
Cox1 Cox2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 0.001 (.) 0.482* (0.26) 0.003** (0.00) 0.316 (0.71)
B. Second stage
VET 0.079*** (0.02) 0.041* (0.02) -2.630 (.) -1.958*** (0.04) -2.627*** (0.01) -1.946*** (0.04)
Age 0.012** (0.01) 0.044*** (0.01) 0.044*** (0.01)
Female 0.009 (0.02) -0.202*** (0.03) -0.202*** (0.03)
Dad: Upp.-sec. -0.007 (0.04) 0.078 (0.05) 0.078 (0.05)
Dad: Post-sec. -0.000 (0.04) 0.026 (0.05) 0.025 (0.05)
Mom: Upp.-sec. 0.037 (0.03) -0.041 (0.04) -0.044 (0.04)
Mom: Post-sec. 0.065* (0.03) -0.177*** (0.05) -0.182*** (0.05)
Marital status -0.039 (0.03) 0.052 (0.05) 0.053 (0.05)
Child -0.050 (0.05) 0.035 (0.08) 0.027 (0.08)
Res: BE 0.010 (0.04) -0.020 (0.06) 0.075 (0.07)
Res: LU -0.051 (0.05) -0.020 (0.06) 0.070 (0.08)
Res: UR -0.072 (0.15) 0.047 (0.19) -0.042 (0.29)
Res: SZ -0.138** (0.06) 0.065 (0.09) 0.166 (0.12)
Res: OW -0.054 (0.13) 0.007 (0.17) 0.185 (0.25)
Res: NW -0.221 (0.16) 0.055 (0.22) 0.241 (0.29)
Res: GL -0.013 (0.12) -0.151 (0.23) -0.172 (0.30)
Res: ZG -0.146** (0.07) 0.167* (0.10) 0.286** (0.12)
Res: FR -0.014 (0.06) 0.034 (0.09) 0.160 (0.10)
Res: SO -0.006 (0.05) 0.022 (0.08) 0.044 (0.09)
Res: BS -0.085 (0.06) -0.050 (0.09) 0.130 (0.10)
Res: BL -0.014 (0.07) -0.144 (0.09) 0.155 (0.11)
Res: SH -0.118 (0.09) 0.183 (0.14) 0.314* (0.18)
Res: AR -0.153 (0.11) 0.320* (0.17) 0.467*** (0.18)
Res: AI -0.245 (0.23) 0.328 (0.30) 0.749** (0.38)
Res: SG -0.003 (0.05) 0.067 (0.07) 0.085 (0.08)
Res: GR 0.027 (0.07) -0.095 (0.10) 0.038 (0.12)
Res: AG -0.002 (0.04) 0.087* (0.05) 0.108* (0.06)
Res: TG -0.099* (0.05) 0.249*** (0.08) 0.344*** (0.10)
Res: TI -0.099 (0.08) 0.028 (0.10) 0.068 (0.15)
Res: VD -0.108* (0.06) 0.059 (0.09) 0.185* (0.10)
Res: VS -0.033 (0.07) -0.046 (0.09) 0.133 (0.12)
Res: NE -0.121 (0.08) 0.001 (0.12) 0.272* (0.14)
Res: GE -0.195** (0.08) -0.020 (0.11) 0.272** (0.14)
Res: JU -0.182** (0.09) 0.237* (0.14) 0.414** (0.19)
UAS -0.154 (0.18) 1.133*** (0.26) 1.075*** (0.26)
Master 0.028 (0.03) -0.246*** (0.04) -0.238*** (0.04)
UNI 2 -0.006 (0.06) 0.012 (0.08) -0.001 (0.08)
UNI 3 0.077 (0.07) -0.035 (0.09) -0.035 (0.09)
UNI 4 0.115 (0.08) 0.025 (0.10) 0.045 (0.10)
UNI 5 -0.003 (0.07) 0.058 (0.09) 0.075 (0.09)
UNI 6 0.104 (0.13) -0.086 (0.16) -0.112 (0.16)
UNI 7 0.053 (0.08) 0.091 (0.11) 0.117 (0.11)
UNI 8 -0.044 (0.07) -0.009 (0.09) -0.020 (0.09)
UNI 9 0.111* (0.06) -0.153** (0.08) -0.193** (0.08)
UNI 10 -0.065 (0.10) 0.099 (0.14) 0.100 (0.14)
UNI 11 0.103 (0.08) -0.316*** (0.10) -0.288*** (0.10)
UNI 12 0.148** (0.06) -0.485*** (0.08) -0.492*** (0.08)
UAS 1 0.248 (0.17) -0.547** (0.26) -0.499** (0.25)
UAS 2 0.252 (0.17) -0.793*** (0.26) -0.713*** (0.25)
UAS 3 0.270 (0.17) -0.537** (0.26) -0.497** (0.25)
UAS 4 0.388** (0.17) -0.591** (0.26) -0.547** (0.25)
UAS 5 0.408** (0.18) -0.742*** (0.27) -0.691** (0.27)
UAS 6 0.344** (0.17) -0.615** (0.26) -0.588** (0.25)
UAS 7 0.298* (0.17) -0.518** (0.25) -0.504** (0.25)
UAS 8 0.000 (.)
Arts -0.124* (0.07) -0.142 (0.09) -0.136 (0.09)
Edu. Sciences 0.047 (0.05) 0.059 (0.07) 0.056 (0.07)
Economics 0.098** (0.05) 0.070 (0.07) 0.067 (0.07)
Nat. Sciences 0.019 (0.05) 0.217*** (0.07) 0.215*** (0.07)
Medi. Sciences 0.213*** (0.08) 0.037 (0.10) 0.019 (0.10)
Health 0.185*** (0.06) -0.079 (0.08) -0.079 (0.08)
Engineering 0.232*** (0.06) 0.218*** (0.08) 0.217*** (0.08)
Agri. Sciences 0.125** (0.05) 0.410*** (0.07) 0.402*** (0.07)
Others -0.165* (0.08) 0.506*** (0.11) 0.485*** (0.11)
No. of semester -0.026*** (0.00) 0.010 (0.01) 0.009 (0.01)
Grade 0.071*** (0.02) -0.036 (0.03) -0.035 (0.03)
Cohort=2013 -0.084** (0.03) -0.040 (0.04) -0.036 (0.04)
Cohort=2015 -0.117*** (0.03) 0.040 (0.04) 0.041 (0.04)
Cohort=2017 -0.098*** (0.03) -0.017 (0.04) -0.012 (0.04)
Pre-study job 0.053** (0.02) -0.085** (0.03) -0.083** (0.03)
(continued on next page)

19
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A6 (continued )
Cox1 Cox2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Student internship 0.031* (0.02) -0.253*** (0.03) -0.248*** (0.03)

Student job -0.047* (0.02) 0.090*** (0.03) 0.086** (0.03)
Work-based high. edu. -0.176*** (0.03) 0.391*** (0.04) 0.379*** (0.04)
Exchange 0.050** (0.02) -0.046 (0.03) -0.044 (0.03)
Public empl. -0.029 (0.02) 0.050* (0.03) 0.047 (0.03)
Medium 0.053** (0.02) -0.052 (0.03) -0.054* (0.03)
Large 0.060*** (0.02) -0.042 (0.03) -0.037 (0.03)
Manufacturing -0.729** (0.28) 0.438 (0.42) 0.459 (0.42)
Electricity ... -0.694** (0.29) 0.419 (0.44) 0.445 (0.44)
Water Supply ... -0.968*** (0.35) 0.636 (0.55) 0.687 (0.54)
Construction -0.922*** (0.29) 0.769* (0.44) 0.784* (0.44)
Wholesale ... -0.752*** (0.28) 0.377 (0.43) 0.399 (0.42)
Transportation ... -0.860*** (0.29) 0.378 (0.43) 0.406 (0.43)
Accommodation ... -1.070*** (0.29) 0.646 (0.44) 0.650 (0.44)
Information ... -0.748*** (0.28) 0.400 (0.42) 0.421 (0.42)
Financial ... -0.735*** (0.28) 0.403 (0.42) 0.427 (0.42)
Real Estate -0.651** (0.30) 0.339 (0.45) 0.368 (0.45)
Scientific ... -0.691** (0.28) 0.306 (0.42) 0.327 (0.42)
Administrative ... -0.766*** (0.29) 0.297 (0.43) 0.321 (0.43)
Public ... -0.813*** (0.29) 0.385 (0.43) 0.406 (0.43)
Education -0.846*** (0.29) 0.446 (0.43) 0.468 (0.43)
Health ... -0.805*** (0.29) 0.321 (0.43) 0.341 (0.43)
Arts ... -0.963*** (0.29) 0.529 (0.43) 0.536 (0.43)
Other service ... -0.861*** (0.29) 0.451 (0.43) 0.462 (0.43)
Extraterritorial ... -0.813*** (0.31) 0.455 (0.45) 0.420 (0.45)
Lausanne -0.065 (0.06) 0.058 (0.08) 0.055 (0.08)
Sion -0.118 (0.09) 0.169 (0.12) 0.164 (0.12)
Fribourg -0.133* (0.08) 0.124 (0.10) 0.124 (0.10)
Neuchatel -0.146* (0.08) 0.246** (0.12) 0.258** (0.12)
Biel -0.145* (0.08) 0.247** (0.11) 0.248** (0.11)
Bern -0.088 (0.07) 0.161* (0.10) 0.162* (0.10)
Basel -0.094 (0.08) 0.055 (0.11) 0.056 (0.11)
Aarau-Olten -0.124 (0.08) 0.124 (0.11) 0.132 (0.11)
Zurich -0.092 (0.07) 0.086 (0.10) 0.091 (0.10)
Winterthur-SH -0.145* (0.08) 0.243** (0.11) 0.250** (0.11)
St. Gallen -0.144* (0.08) 0.184 (0.12) 0.187 (0.12)
Chur -0.197* (0.10) 0.162 (0.14) 0.155 (0.14)
Luzern -0.086 (0.08) 0.105 (0.11) 0.115 (0.11)
Bellinzona -0.216* (0.12) 0.252 (0.16) 0.260 (0.16)
Lugano -0.299*** (0.11) 0.237 (0.15) 0.242* (0.15)
Abroad -0.359*** (0.10) 0.392** (0.16) 0.397** (0.16)
Ori: BE -0.162*** (0.05) -0.178*** (0.06)
Ori: LU -0.282*** (0.05) -0.180** (0.07)
Ori: UR -0.482*** (0.05) 0.011 (0.24)
Ori: SZ -0.396*** (0.14) -0.200* (0.11)
Ori: OW -0.436*** (0.05) -0.279 (0.21)
Ori: NW -0.358*** (0.05) -0.282 (0.21)
Ori: GL -0.286*** (0.05) -0.034 (0.25)
Ori: ZG -0.234*** (0.05) -0.226* (0.13)
Ori: FR -0.427*** (0.05) -0.233*** (0.09)
Ori: SO -0.018 (0.11) -0.071 (0.08)
Ori: BS -0.728*** (0.05) -0.266** (0.11)
Ori: BL -0.627*** (0.05) -0.437*** (0.09)
Ori: SH 0.222*** (0.05) -0.221 (0.15)
Ori: AR -0.158*** (0.05) -0.253* (0.13)
Ori: AI -0.731*** (0.05) -0.648* (0.35)
Ori: SG 0.011 (0.09) -0.068 (0.07)
Ori: GR -0.373*** (0.12) -0.224** (0.10)
Ori: AG -0.016 (0.08) -0.087 (0.06)
Ori: TG 0.115** (0.05) -0.175** (0.08)
Ori: TI -0.509*** (0.05) -0.141 (0.12)
Ori: VD -0.642*** (0.05) -0.222*** (0.08)
Ori: VS -0.582*** (0.05) -0.289*** (0.09)
Ori: NE -0.557*** (0.05) -0.414*** (0.12)
Ori: GE -1.060*** (0.05) -0.430*** (0.11)
Ori: JU -0.259*** (0.05) -0.300** (0.15)
Constant 1.531 (.) 0.319 (0.53) 1.889*** (0.05) 0.403 (0.53)
Observations 12,075 12,075 13,571 13,571 13,571 13,571
Pseudo R2 0.000 0.002

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the Cox regressions (first two columns) and IV regressions (last four columns) on
search time (ln). The data is from the one-year post-graduation surveys. Columns Cox1 and IV1 include no controls, whereas IV3 includes region of origin fixed effects.
Columns Cox2 and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first stage include the same
controls as the estimations from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance at the 1%, 5% and 10%
level, respectively.

20
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A7
Regression table for internship position one year post-graduation.
Probit1 Probit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 2.982*** (0.11) 1.365*** (0.34) 3.219*** (0.65) 0.751 (1.04)
B. Second stage
VET -0.559*** (0.03) -0.238*** (0.04) -0.631*** (0.11) -0.043 (0.15) -1.701*** (0.13) -0.110 (0.15)
Age -0.053*** (0.01) -0.058*** (0.01) -0.056*** (0.01)
Female 0.184*** (0.03) 0.201*** (0.03) 0.197*** (0.04)
Dad: Upp.-sec. 0.058 (0.07) 0.052 (0.07) 0.049 (0.07)
Dad: Post-sec. 0.036 (0.07) 0.034 (0.07) 0.033 (0.07)
Mom: Upp.-sec. -0.055 (0.06) -0.056 (0.06) -0.055 (0.06)
Mom: Post-sec. -0.004 (0.06) 0.004 (0.06) 0.001 (0.06)
Marital status -0.224*** (0.08) -0.223*** (0.08) -0.226*** (0.08)
Child -0.255* (0.14) -0.252* (0.14) -0.257* (0.14)
Res: BE 0.068 (0.08) 0.069 (0.08) 0.102 (0.09)
Res: LU 0.098 (0.08) 0.106 (0.08) 0.084 (0.11)
Res: UR 0.039 (0.25) 0.042 (0.25) 0.079 (0.41)
Res: SZ -0.016 (0.13) -0.007 (0.13) -0.032 (0.15)
Res: OW -0.013 (0.25) -0.003 (0.25) -0.111 (0.32)
Res: NW 0.216 (0.23) 0.231 (0.23) 0.226 (0.32)
Res: GL -0.238 (0.36) -0.217 (0.36) -0.401 (0.43)
Res: ZG -0.023 (0.14) -0.024 (0.14) 0.073 (0.18)
Res: FR -0.003 (0.11) -0.004 (0.11) 0.144 (0.14)
Res: SO 0.035 (0.11) 0.035 (0.11) 0.108 (0.13)
Res: BS -0.152 (0.12) -0.138 (0.12) -0.094 (0.13)
Res: BL -0.190 (0.12) -0.175 (0.12) -0.142 (0.15)
Res: SH -0.032 (0.18) -0.039 (0.18) 0.404 (0.27)
Res: AR 0.102 (0.23) 0.090 (0.23) -0.019 (0.27)
Res: AI -0.002 (0.40) -0.007 (0.40) 0.020 (0.56)
Res: SG 0.088 (0.09) 0.081 (0.09) 0.070 (0.11)
Res: GR 0.263** (0.13) 0.270** (0.13) 0.349** (0.16)
Res: AG 0.046 (0.07) 0.038 (0.07) 0.107 (0.09)
Res: TG 0.024 (0.11) 0.015 (0.11) 0.076 (0.14)
Res: TI 0.416*** (0.11) 0.423*** (0.11) 0.407** (0.20)
Res: VD 0.004 (0.11) 0.007 (0.11) 0.103 (0.13)
Res: VS -0.126 (0.12) -0.116 (0.13) -0.098 (0.17)
Res: NE 0.020 (0.14) 0.027 (0.14) 0.067 (0.18)
Res: GE 0.231* (0.14) 0.250* (0.14) 0.471*** (0.18)
Res: JU -0.183 (0.22) -0.184 (0.22) -0.124 (0.28)
UAS -1.129** (0.46) -1.238*** (0.46) -1.211*** (0.46)
Master -0.236*** (0.06) -0.214*** (0.06) -0.216*** (0.06)
UNI 2 -0.163 (0.11) -0.165 (0.11) -0.165 (0.11)
UNI 3 -0.444*** (0.12) -0.447*** (0.12) -0.440*** (0.12)
UNI 4 -0.168 (0.13) -0.186 (0.13) -0.154 (0.14)
UNI 5 -0.043 (0.12) -0.049 (0.12) -0.034 (0.12)
UNI 6 -0.031 (0.20) -0.036 (0.20) -0.038 (0.20)
UNI 7 -0.232* (0.14) -0.246* (0.14) -0.234 (0.14)
UNI 8 -0.799*** (0.13) -0.795*** (0.13) -0.809*** (0.13)
UNI 9 -0.203** (0.10) -0.201* (0.10) -0.219** (0.10)
UNI 10 -0.477*** (0.18) -0.475*** (0.18) -0.477*** (0.18)
UNI 11 -0.231* (0.13) -0.215* (0.13) -0.193 (0.13)
UNI 12 -0.531*** (0.11) -0.504*** (0.11) -0.522*** (0.11)
UAS 1 0.231 (0.45) 0.261 (0.45) 0.263 (0.45)
UAS 2 0.332 (0.45) 0.388 (0.45) 0.399 (0.45)
UAS 3 0.321 (0.45) 0.348 (0.45) 0.355 (0.45)
UAS 4 0.281 (0.45) 0.299 (0.45) 0.295 (0.45)
UAS 5 0.232 (0.47) 0.274 (0.47) 0.272 (0.47)
UAS 6 0.403 (0.45) 0.430 (0.45) 0.419 (0.45)
UAS 7 0.363 (0.44) 0.386 (0.45) 0.375 (0.45)
UAS 8 0.000 (.)
Arts 0.195* (0.10) 0.221** (0.11) 0.214** (0.11)
Edu. Sciences -0.221*** (0.08) -0.234*** (0.08) -0.234*** (0.08)
Economics -0.113 (0.08) -0.132* (0.08) -0.128 (0.08)
Nat. Sciences -0.066 (0.09) -0.089 (0.09) -0.081 (0.09)
Medi. Sciences -1.390*** (0.17) -1.416*** (0.18) -1.414*** (0.18)
Health -1.248*** (0.13) -1.252*** (0.13) -1.250*** (0.13)
Engineering -0.732*** (0.10) -0.776*** (0.11) -0.762*** (0.11)
Agri. Sciences -0.377*** (0.09) -0.428*** (0.10) -0.413*** (0.10)
Others -0.115 (0.13) -0.142 (0.13) -0.147 (0.13)
No. of semester -0.032*** (0.01) -0.030*** (0.01) -0.032*** (0.01)
Grade -0.151*** (0.04) -0.154*** (0.04) -0.154*** (0.04)
Cohort=2013 0.153*** (0.05) 0.169*** (0.05) 0.164*** (0.05)
Cohort=2015 0.198*** (0.05) 0.210*** (0.05) 0.208*** (0.05)
Cohort=2017 0.239*** (0.05) 0.255*** (0.05) 0.251*** (0.05)
Pre-study job -0.011 (0.05) -0.009 (0.05) -0.006 (0.05)
(continued on next page)

21
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A7 (continued )
Probit1 Probit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Student internship -0.208*** (0.03) -0.188*** (0.04) -0.195*** (0.04)

Student job 0.010 (0.05) 0.007 (0.05) 0.009 (0.05)
Work-based high. edu. -0.647*** (0.07) -0.660*** (0.07) -0.661*** (0.07)
Exchange 0.164*** (0.04) 0.164*** (0.04) 0.166*** (0.04)
Public empl. -0.002 (0.04) -0.003 (0.04) -0.004 (0.04)
Medium -0.053 (0.04) -0.053 (0.04) -0.054 (0.04)
Large 0.061 (0.04) 0.060 (0.04) 0.060 (0.04)
Manufacturing 0.047 (0.51) 0.086 (0.53) 0.071 (0.52)
Electricity ... 0.183 (0.53) 0.221 (0.55) 0.199 (0.54)
Water Supply ... 0.336 (0.60) 0.382 (0.61) 0.363 (0.61)
Construction 0.099 (0.53) 0.129 (0.55) 0.119 (0.54)
Wholesale ... 0.131 (0.51) 0.178 (0.53) 0.159 (0.53)
Transportation ... 0.336 (0.52) 0.395 (0.54) 0.377 (0.53)
Accommodation ... 0.054 (0.53) 0.119 (0.54) 0.090 (0.54)
Information ... 0.150 (0.51) 0.194 (0.53) 0.181 (0.52)
Financial ... 0.017 (0.51) 0.063 (0.53) 0.049 (0.53)
Real Estate 0.102 (0.54) 0.146 (0.56) 0.129 (0.55)
Scientific ... 0.140 (0.51) 0.187 (0.53) 0.171 (0.52)
Administrative ... 0.169 (0.52) 0.223 (0.54) 0.205 (0.53)
Public ... 0.575 (0.51) 0.628 (0.53) 0.612 (0.53)
Education -0.256 (0.52) -0.203 (0.53) -0.222 (0.53)
Health ... -0.035 (0.51) 0.022 (0.53) -0.004 (0.53)
Arts ... 0.213 (0.52) 0.270 (0.54) 0.245 (0.53)
Other service ... 0.534 (0.52) 0.588 (0.53) 0.559 (0.53)
Extraterritorial ... 0.520 (0.58) 0.575 (0.59) 0.505 (0.59)
Lausanne 0.251** (0.10) 0.253** (0.10) 0.248** (0.10)
Sion 0.199 (0.16) 0.192 (0.16) 0.190 (0.16)
Fribourg 0.346** (0.14) 0.347** (0.14) 0.352*** (0.14)
Neuchatel 0.214 (0.15) 0.206 (0.15) 0.202 (0.15)
Biel 0.154 (0.15) 0.147 (0.15) 0.147 (0.15)
Bern 0.440*** (0.12) 0.433*** (0.12) 0.431*** (0.12)
Basel 0.458*** (0.14) 0.460*** (0.14) 0.454*** (0.14)
Aarau-Olten 0.193 (0.14) 0.197 (0.14) 0.195 (0.14)
Zurich 0.287** (0.12) 0.287** (0.12) 0.285** (0.12)
Winterthur-SH 0.145 (0.15) 0.136 (0.15) 0.128 (0.15)
St. Gallen 0.072 (0.15) 0.071 (0.15) 0.067 (0.15)
Chur 0.068 (0.18) 0.071 (0.18) 0.063 (0.18)
Luzern 0.128 (0.14) 0.129 (0.14) 0.121 (0.14)
Bellinzona 0.141 (0.20) 0.138 (0.20) 0.140 (0.20)
Lugano -0.049 (0.18) -0.044 (0.18) -0.046 (0.18)
Abroad 0.729*** (0.16) 0.730*** (0.16) 0.733*** (0.16)
Ori: BE -0.040 (0.05) -0.073 (0.08)
Ori: LU -0.157*** (0.06) -0.006 (0.10)
Ori: UR -0.306* (0.18) -0.076 (0.33)
Ori: SZ -0.226** (0.09) -0.015 (0.13)
Ori: OW -0.113 (0.15) 0.080 (0.26)
Ori: NW -0.150 (0.17) -0.040 (0.28)
Ori: GL -0.083 (0.21) 0.162 (0.30)
Ori: ZG -0.218** (0.11) -0.175 (0.17)
Ori: FR -0.305*** (0.07) -0.239* (0.13)
Ori: SO -0.011 (0.07) -0.129 (0.11)
Ori: BS -0.332*** (0.13) -0.092 (0.14)
Ori: BL -0.260*** (0.08) -0.082 (0.12)
Ori: SH -0.096 (0.13) -0.564** (0.24)
Ori: AR 0.037 (0.13) 0.103 (0.19)
Ori: AI -0.418 (0.30) -0.082 (0.46)
Ori: SG -0.026 (0.06) -0.017 (0.09)
Ori: GR -0.081 (0.07) -0.129 (0.13)
Ori: AG 0.016 (0.05) -0.119 (0.09)
Ori: TG -0.075 (0.08) -0.100 (0.11)
Ori: TI -0.312*** (0.09) -0.036 (0.17)
Ori: VD -0.390*** (0.07) -0.160 (0.12)
Ori: VS -0.426*** (0.07) -0.070 (0.13)
Ori: NE -0.361*** (0.09) -0.092 (0.15)
Ori: GE -0.600*** (0.10) -0.325** (0.15)
Ori: JU -0.320*** (0.10) -0.120 (0.20)
Constant -0.794*** (0.02) 1.784*** (0.65) -0.765*** (0.05) 1.841*** (0.67) 0.030 (0.13) 1.868*** (0.66)
Observations 13,571 13,571 13,571 13,571 13,571 13,571
Pseudo R2 0.033 0.178

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the probit regressions (first two columns) and IV regressions (last four columns)
on internship. The data is from the one-year post-graduation surveys. Columns Probit1 and IV1 include no controls, whereas IV3 includes region of origin fixed effects.
Columns Probit2 and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first stage include the same
controls as the estimations from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance at the 1%, 5% and 10%
level, respectively.

22
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A8
Regression table for wage five years post-graduation.
Tobit1 Tobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 2.539*** (0.24) 1.176 (0.76) -2.234 (2.29) -3.315 (2.90)
B. Second stage
VET 0.085*** (0.01) 0.027 (0.02) 0.228*** (0.04) 0.118** (0.05) 0.143*** (0.05) 0.117** (0.05)
Age 0.016*** (0.01) 0.014** (0.01) 0.014** (0.01)
Female -0.100*** (0.01) -0.090*** (0.02) -0.090*** (0.02)
Marital status 0.014 (0.01) 0.012 (0.01) 0.012 (0.01)
Child -0.168*** (0.02) -0.167*** (0.02) -0.170*** (0.02)
Dad: Upp.-sec. -0.039 (0.02) -0.042* (0.02) -0.040 (0.02)
Dad: Post-sec. -0.026 (0.02) -0.029 (0.02) -0.025 (0.02)
Mom: Upp.-sec. 0.016 (0.02) 0.015 (0.02) 0.015 (0.02)
Mom: Post-sec. 0.002 (0.03) 0.004 (0.03) 0.005 (0.03)
Res: BE -0.094*** (0.04) -0.096*** (0.04) -0.092** (0.04)
Res: LU -0.014 (0.04) -0.008 (0.04) -0.029 (0.04)
Res: UR 0.135 (0.12) 0.138 (0.12) 0.345** (0.15)
Res: SZ 0.026 (0.05) 0.028 (0.05) -0.017 (0.07)
Res: OW 0.061 (0.11) 0.054 (0.11) -0.102 (0.12)
Res: NW 0.049 (0.10) 0.055 (0.10) 0.096 (0.10)
Res: GL -0.101** (0.05) -0.100** (0.05) 0.008 (0.09)
Res: ZG -0.014 (0.04) -0.007 (0.04) 0.025 (0.04)
Res: FR -0.084** (0.04) -0.084** (0.04) -0.040 (0.05)
Res: SO -0.064* (0.04) -0.074* (0.04) -0.085* (0.05)
Res: BS -0.096** (0.05) -0.084* (0.05) -0.088* (0.05)
Res: BL 0.003 (0.05) 0.011 (0.05) 0.005 (0.06)
Res: SH 0.080 (0.06) 0.074 (0.06) 0.131 (0.10)
Res: AR 0.115* (0.07) 0.113* (0.07) 0.141 (0.09)
Res: AI -0.512 (0.45) -0.516 (0.46) -0.649 (0.46)
Res: SG -0.003 (0.03) -0.009 (0.03) 0.001 (0.04)
Res: GR 0.041 (0.08) 0.052 (0.08) 0.069 (0.08)
Res: AG -0.001 (0.03) -0.002 (0.03) 0.021 (0.03)
Res: TG 0.034 (0.04) 0.020 (0.04) 0.002 (0.05)
Res: TI -0.032 (0.06) -0.029 (0.06) -0.076 (0.07)
Res: VD -0.066 (0.04) -0.062 (0.04) -0.037 (0.05)
Res: VS -0.055 (0.06) -0.053 (0.06) -0.021 (0.07)
Res: NE -0.085 (0.06) -0.080 (0.06) -0.029 (0.06)
Res: GE -0.138** (0.06) -0.129** (0.06) -0.088 (0.07)
Res: JU -0.148** (0.07) -0.158** (0.07) -0.160 (0.12)
UAS -0.060 (0.29) -0.124 (0.30) -0.116 (0.28)
Master 0.116*** (0.03) 0.121*** (0.03) 0.117*** (0.03)
UNI 2 0.038 (0.05) 0.037 (0.05) 0.036 (0.05)
UNI 3 0.036 (0.05) 0.030 (0.05) 0.028 (0.05)
UNI 4 0.057 (0.06) 0.051 (0.06) 0.064 (0.06)
UNI 5 0.053 (0.05) 0.051 (0.05) 0.056 (0.05)
UNI 6 -0.056 (0.15) -0.065 (0.15) -0.070 (0.15)
UNI 7 -0.001 (0.06) -0.009 (0.06) 0.004 (0.06)
UNI 8 0.159*** (0.06) 0.166*** (0.06) 0.164*** (0.06)
UNI 9 -0.014 (0.05) -0.015 (0.05) -0.019 (0.05)
UNI 10 0.002 (0.07) 0.010 (0.07) -0.002 (0.07)
UNI 11 0.016 (0.05) 0.029 (0.05) 0.036 (0.05)
UNI 12 0.045 (0.04) 0.059 (0.04) 0.054 (0.05)
UAS 1 0.074 (0.28) 0.094 (0.29) 0.076 (0.28)
UAS 2 0.105 (0.28) 0.136 (0.29) 0.131 (0.28)
UAS 3 0.086 (0.28) 0.101 (0.29) 0.090 (0.28)
UAS 4 0.054 (0.28) 0.070 (0.29) 0.047 (0.28)
UAS 5 0.114 (0.29) 0.130 (0.29) 0.110 (0.28)
UAS 6 0.061 (0.28) 0.078 (0.29) 0.065 (0.28)
UAS 7 0.091 (0.28) 0.110 (0.29) 0.092 (0.28)
UAS 8 0.000 (.)
Arts -0.119 (0.07) -0.109 (0.07) -0.111 (0.07)
Edu. Sciences 0.109** (0.04) 0.103** (0.04) 0.098** (0.04)
Economics 0.276*** (0.04) 0.261*** (0.04) 0.259*** (0.04)
Nat. Sciences 0.153*** (0.05) 0.139*** (0.05) 0.139*** (0.05)
Medi. Sciences 0.214*** (0.07) 0.203*** (0.08) 0.201*** (0.07)
Health 0.057 (0.05) 0.056 (0.05) 0.056 (0.05)
Engineering 0.192*** (0.05) 0.166*** (0.05) 0.166*** (0.05)
Agri. Sciences 0.256*** (0.05) 0.227*** (0.05) 0.228*** (0.05)
Others 0.011 (0.09) -0.004 (0.09) -0.006 (0.09)
Grade 0.026 (0.02) 0.025 (0.02) 0.024 (0.02)
Cohort=2017 -0.022 (0.01) -0.014 (0.01) -0.015 (0.01)
Exchange -0.001 (0.02) -0.002 (0.02) 0.000 (0.02)
Pre-study job 0.026 (0.02) 0.028 (0.02) 0.029* (0.02)
Student internship 0.012 (0.01) 0.021 (0.01) 0.022 (0.01)
Student job 0.032* (0.02) 0.030* (0.02) 0.029* (0.02)
(continued on next page)

23
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A8 (continued )
Tobit1 Tobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Work-based high. edu. 0.057** (0.02) 0.048* (0.03) 0.052** (0.03)

Cont. studies -0.195*** (0.03) -0.196*** (0.03) -0.200*** (0.03)
Add. studies 0.010 (0.02) 0.008 (0.02) 0.009 (0.02)
Cont. education 0.046*** (0.02) 0.046*** (0.02) 0.047*** (0.02)
Medium 0.064*** (0.02) 0.063*** (0.02) 0.062*** (0.02)
Large 0.122*** (0.02) 0.122*** (0.02) 0.121*** (0.02)
Public empl. -0.009 (0.02) -0.007 (0.02) -0.004 (0.02)
Electricity ... -0.004 (0.06) -0.008 (0.06) -0.021 (0.06)
Water Supply ... -0.466 (0.30) -0.428 (0.29) -0.429 (0.30)
Construction 0.134*** (0.05) 0.129*** (0.05) 0.148*** (0.05)
Wholesale ... 0.050* (0.03) 0.041 (0.03) 0.042 (0.03)
Transportation ... -0.079 (0.05) -0.074 (0.05) -0.066 (0.05)
Accommodation ... -0.185** (0.09) -0.180** (0.08) -0.171** (0.09)
Information ... 0.033 (0.02) 0.033 (0.02) 0.038 (0.02)
Financial ... 0.146*** (0.02) 0.145*** (0.02) 0.148*** (0.02)
Real Estate 0.030 (0.16) 0.033 (0.15) 0.038 (0.15)
Scientific ... 0.040** (0.02) 0.041** (0.02) 0.044** (0.02)
Administrative ... -0.121* (0.07) -0.114 (0.07) -0.109 (0.07)
Public ... 0.037 (0.03) 0.037 (0.03) 0.040 (0.03)
Education -0.190*** (0.04) -0.185*** (0.04) -0.185*** (0.04)
Health ... -0.091*** (0.03) -0.086*** (0.03) -0.085*** (0.03)
Arts ... -0.139** (0.06) -0.133** (0.06) -0.129** (0.06)
Other service ... -0.141*** (0.05) -0.143*** (0.05) -0.140*** (0.05)
Extraterritorial ... -0.187 (0.20) -0.184 (0.21) -0.175 (0.21)
Lausanne -0.084** (0.04) -0.081** (0.04) -0.083** (0.04)
Sion -0.109* (0.06) -0.108* (0.06) -0.107* (0.06)
Fribourg -0.062 (0.05) -0.065 (0.05) -0.065 (0.05)
Neuchatel -0.099* (0.06) -0.109* (0.06) -0.111* (0.06)
Biel -0.105* (0.06) -0.102* (0.06) -0.101 (0.06)
Bern -0.045 (0.05) -0.041 (0.05) -0.041 (0.05)
Basel -0.033 (0.06) -0.031 (0.06) -0.026 (0.06)
Aarau-Olten -0.110* (0.06) -0.103* (0.06) -0.101* (0.06)
Zurich -0.027 (0.05) -0.023 (0.05) -0.021 (0.05)
Winterthur-SH -0.109* (0.06) -0.104* (0.06) -0.107* (0.06)
St. Gallen -0.080 (0.06) -0.072 (0.06) -0.073 (0.06)
Chur -0.139 (0.09) -0.150* (0.09) -0.148* (0.09)
Luzern -0.123* (0.06) -0.121* (0.06) -0.121* (0.07)
Bellinzona -0.169** (0.08) -0.167** (0.08) -0.163** (0.08)
Lugano -0.204** (0.08) -0.207** (0.08) -0.200** (0.08)
Abroad -0.142 (0.10) -0.136 (0.10) -0.132 (0.10)
Ori: BE -0.095*** (0.03) -0.001 (0.03)
Ori: LU -0.017 (0.04) 0.021 (0.04)
Ori: UR -0.110 (0.09) -0.232** (0.11)
Ori: SZ 0.085* (0.04) 0.074 (0.06)
Ori: OW 0.078 (0.07) 0.183*** (0.07)
Ori: NW 0.002 (0.04) -0.056 (0.07)
Ori: GL -0.069 (0.07) -0.141 (0.09)
Ori: ZG -0.027 (0.08) -0.076 (0.06)
Ori: FR -0.060* (0.03) -0.062 (0.04)
Ori: SO -0.025 (0.04) 0.023 (0.05)
Ori: BS -0.040 (0.10) -0.001 (0.07)
Ori: BL -0.026 (0.05) -0.008 (0.05)
Ori: SH 0.025 (0.07) -0.064 (0.08)
Ori: AR -0.037 (0.08) -0.030 (0.07)
Ori: AI -0.321 (0.32) 0.130** (0.05)
Ori: SG -0.031 (0.03) -0.017 (0.03)
Ori: GR -0.045 (0.06) -0.032 (0.05)
Ori: AG -0.025 (0.03) -0.043 (0.04)
Ori: TG -0.014 (0.05) 0.035 (0.05)
Ori: TI -0.084** (0.04) 0.042 (0.04)
Ori: VD -0.032 (0.03) -0.034 (0.04)
Ori: VS -0.072** (0.03) -0.043 (0.05)
Ori: NE -0.117*** (0.04) -0.077 (0.05)
Ori: GE -0.050 (0.04) -0.060 (0.05)
Ori: JU -0.199*** (0.05) -0.009 (0.10)
Constant 11.288*** (0.01) 10.533*** (0.22) 11.231*** (0.02) 10.582*** (0.22) 11.306*** (0.03) 10.592*** (0.22)
Observations 2817 2817 2817 2817 2817 2817
Pseudo R2 0.013 0.548

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the tobit regressions (first two columns) and IV regressions (last four columns)
on wage (ln). The data is from the five-year post-graduation surveys. Columns Tobit1 and IV1 include no controls, whereas IV3 includes region of origin fixed effects.
Columns Tobit2 and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first stage include the same
controls as the estimations from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance at the 1%, 5% and 10%
level, respectively.

24
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A9
Regression table for unemployment during the five years post-graduation.
Probit1 Probit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 2.663*** (0.22) 1.169 (0.80) -2.836 (2.00) -3.249 (2.93)
B. Second stage
VET -0.242*** (0.06) 0.015 (0.09) -1.549*** (0.05) 0.024 (0.38) -0.966* (0.58) -0.049 (0.41)
Age 0.086*** (0.03) 0.086*** (0.03) 0.096*** (0.03)
Female 0.092 (0.07) 0.093 (0.08) 0.086 (0.08)
Marital status -0.056 (0.09) -0.056 (0.09) -0.071 (0.09)
Child -0.075 (0.11) -0.074 (0.11) -0.058 (0.12)
Dad: Upp.-sec. -0.096 (0.13) -0.097 (0.13) -0.085 (0.13)
Dad: Post-sec. -0.236* (0.13) -0.237* (0.13) -0.236* (0.13)
Mom: Upp.-sec. -0.196* (0.11) -0.196* (0.11) -0.198* (0.11)
Mom: Post-sec. -0.117 (0.12) -0.117 (0.12) -0.138 (0.12)
Res: BE 0.361** (0.18) 0.361** (0.18) 0.398** (0.20)
Res: LU 0.128 (0.21) 0.129 (0.21) 0.042 (0.26)
Res: UR 0.000 (.) 0.000 (.) 0.000 (.)
Res: SZ -0.225 (0.30) -0.225 (0.30) 0.264 (0.30)
Res: OW 0.000 (.) 0.000 (.) 0.000 (.)
Res: NW 0.000 (.) 0.000 (.) 0.000 (.)
Res: GL 0.000 (.) 0.000 (.) 0.000 (.)
Res: ZG 0.044 (0.22) 0.045 (0.22) -0.035 (0.26)
Res: FR 0.306 (0.25) 0.306 (0.25) 0.466 (0.30)
Res: SO 0.100 (0.23) 0.099 (0.23) 0.114 (0.28)
Res: BS 0.456* (0.26) 0.457* (0.26) 0.454* (0.27)
Res: BL 0.173 (0.29) 0.173 (0.29) 0.207 (0.32)
Res: SH -0.000 (0.37) -0.001 (0.37) -0.363 (0.48)
Res: AR 0.000 (.) 0.000 (.) 0.000 (.)
Res: AI 0.000 (.) 0.000 (.) 0.000 (.)
Res: SG 0.147 (0.19) 0.146 (0.19) -0.011 (0.22)
Res: GR 0.051 (0.34) 0.052 (0.34) 0.051 (0.41)
Res: AG -0.058 (0.15) -0.058 (0.15) 0.149 (0.18)
Res: TG -0.423 (0.31) -0.425 (0.31) -0.449 (0.33)
Res: TI 0.367 (0.30) 0.367 (0.30) 0.365 (0.36)
Res: VD 0.224 (0.23) 0.224 (0.23) 0.065 (0.27)
Res: VS -0.305 (0.30) -0.305 (0.30) -0.045 (0.34)
Res: NE 0.006 (0.27) 0.006 (0.27) -0.351 (0.32)
Res: GE 0.114 (0.29) 0.115 (0.30) -0.206 (0.36)
Res: JU 0.284 (0.34) 0.283 (0.34) -0.050 (0.48)
UAS -0.159 (0.25) -3.768 (.) -3.558*** (0.51)
Master 0.102 (0.12) 0.103 (0.12) 0.091 (0.12)
UNI 2 -0.166 (0.25) -0.166 (0.25) -0.174 (0.26)
UNI 3 -0.105 (0.27) -0.105 (0.27) -0.097 (0.28)
UNI 4 0.066 (0.28) 0.065 (0.28) -0.022 (0.30)
UNI 5 0.178 (0.26) 0.178 (0.26) 0.090 (0.27)
UNI 6 0.407 (0.47) 0.406 (0.47) 0.516 (0.48)
UNI 7 0.362 (0.29) 0.361 (0.29) 0.211 (0.31)
UNI 8 -0.582* (0.34) -0.581* (0.34) -0.596 (0.36)
UNI 9 -0.263 (0.25) -0.263 (0.25) -0.256 (0.26)
UNI 10 0.065 (0.37) 0.066 (0.37) 0.009 (0.39)
UNI 11 0.272 (0.27) 0.274 (0.27) 0.188 (0.29)
UNI 12 -0.291 (0.25) -0.290 (0.26) -0.333 (0.27)
UAS 1 -0.121 (0.18) 3.484 (.) 3.317*** (0.37)
UAS 2 0.172 (0.16) 3.778 (.) 3.531*** (0.38)
UAS 3 -0.079 (0.16) 3.525 (.) 3.379*** (0.35)
UAS 4 0.008 (0.17) 3.612 (.) 3.430*** (0.35)
UAS 5 0.232 (0.29) 3.836 (.) 3.640*** (0.43)
UAS 6 -0.308* (0.17) 3.297 (.) 3.077*** (0.35)
UAS 7 0.000 (.) 3.604 (.) 3.420*** (0.32)
UAS 8 0.000 (.)
Arts 0.248 (0.24) 0.249 (0.24) 0.263 (0.25)
Edu. Sciences -0.319* (0.17) -0.319* (0.17) -0.257 (0.18)
Economics -0.306* (0.17) -0.307* (0.18) -0.274 (0.18)
Nat. Sciences -0.046 (0.19) -0.048 (0.20) 0.000 (0.20)
Medi. Sciences -1.969*** (0.45) -1.970*** (0.45) -1.938*** (0.46)
Health -0.947*** (0.24) -0.947*** (0.24) -0.920*** (0.24)
Engineering -0.894*** (0.23) -0.897*** (0.25) -0.921*** (0.25)
Agri. Sciences -0.512*** (0.19) -0.515** (0.22) -0.480** (0.23)
Others -0.280 (0.31) -0.281 (0.31) -0.243 (0.31)
Grade -0.487*** (0.10) -0.487*** (0.10) -0.477*** (0.10)
Cohort=2017 0.087 (0.07) 0.088 (0.08) 0.069 (0.08)
Exchange 0.193** (0.08) 0.193** (0.08) 0.202** (0.08)
Pre-study job -0.245*** (0.09) -0.245*** (0.09) -0.255*** (0.09)
Student internship -0.130* (0.07) -0.129 (0.08) -0.128 (0.08)
Student job -0.124 (0.10) -0.124 (0.10) -0.133 (0.10)
(continued on next page)

25
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A9 (continued )
Probit1 Probit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Work-based high. edu. -0.335*** (0.13) -0.336** (0.14) -0.325** (0.14)

Cont. studies -0.075 (0.10) -0.076 (0.10) -0.065 (0.10)
Add. studies -0.177* (0.10) -0.177* (0.10) -0.161 (0.10)
Cont. education -0.176* (0.11) -0.176* (0.11) -0.176* (0.11)
Medium 0.018 (0.09) 0.018 (0.09) 0.028 (0.09)
Large -0.175** (0.08) -0.175** (0.08) -0.178** (0.08)
Public empl. -0.132 (0.10) -0.132 (0.10) -0.134 (0.10)
Electricity ... 0.088 (0.29) 0.088 (0.29) 0.133 (0.30)
Water Supply ... 0.000 (.) 0.000 (.) 0.000 (.)
Construction -0.111 (0.30) -0.111 (0.30) -0.055 (0.30)
Wholesale ... 0.044 (0.14) 0.043 (0.14) 0.056 (0.15)
Transportation ... -0.066 (0.21) -0.066 (0.21) -0.113 (0.21)
Accommodation ... -0.051 (0.39) -0.050 (0.39) -0.023 (0.39)
Information ... -0.417*** (0.13) -0.417*** (0.13) -0.395*** (0.14)
Financial ... -0.184 (0.15) -0.184 (0.15) -0.174 (0.15)
Real Estate -0.456 (0.40) -0.456 (0.40) -0.544 (0.41)
Scientific ... -0.281** (0.12) -0.281** (0.12) -0.270** (0.12)
Administrative ... 0.066 (0.21) 0.066 (0.21) 0.019 (0.21)
Public ... -0.031 (0.18) -0.031 (0.18) -0.048 (0.18)
Education 0.161 (0.19) 0.162 (0.19) 0.138 (0.19)
Health ... 0.127 (0.16) 0.127 (0.16) 0.106 (0.16)
Arts ... 0.045 (0.23) 0.046 (0.24) 0.045 (0.24)
Other service ... -0.448** (0.21) -0.448** (0.21) -0.473** (0.21)
Extraterritorial ... -0.573 (0.70) -0.572 (0.70) -0.453 (0.74)
Lausanne -0.176 (0.19) -0.176 (0.19) -0.136 (0.19)
Sion -0.202 (0.34) -0.202 (0.34) -0.129 (0.35)
Fribourg -0.474* (0.27) -0.474* (0.27) -0.365 (0.28)
Neuchatel -0.064 (0.27) -0.065 (0.28) -0.051 (0.28)
Biel -0.367 (0.28) -0.367 (0.28) -0.310 (0.29)
Bern -0.557** (0.25) -0.557** (0.25) -0.463* (0.25)
Basel -0.627** (0.32) -0.626** (0.32) -0.545* (0.32)
Aarau-Olten -0.525* (0.29) -0.524* (0.29) -0.442 (0.30)
Zurich -0.258 (0.25) -0.257 (0.26) -0.176 (0.26)
Winterthur-SH -0.601* (0.32) -0.601* (0.32) -0.508 (0.33)
St. Gallen -0.600* (0.33) -0.599* (0.33) -0.558* (0.33)
Chur -0.463 (0.41) -0.465 (0.41) -0.385 (0.42)
Luzern -0.572* (0.33) -0.572* (0.33) -0.455 (0.34)
Bellinzona -0.428 (0.42) -0.427 (0.42) -0.329 (0.42)
Lugano -0.213 (0.39) -0.213 (0.39) -0.160 (0.39)
Abroad -0.355 (0.50) -0.355 (0.50) -0.249 (0.51)
Ori: BE 0.047 (0.13) -0.046 (0.17)
Ori: LU -0.088 (0.17) 0.134 (0.23)
Ori: UR 0.000 (.) 0.000 (.)
Ori: SZ -0.717** (0.30) -0.852** (0.33)
Ori: OW 0.000 (.) 0.000 (.)
Ori: NW -0.137 (0.43) 0.311 (0.53)
Ori: GL 0.000 (.) 0.000 (.)
Ori: ZG -0.036 (0.28) 0.208 (0.34)
Ori: FR 0.156 (0.16) -0.236 (0.26)
Ori: SO 0.139 (0.17) 0.078 (0.24)
Ori: BS -0.036 (0.39) 0.069 (0.32)
Ori: BL -0.185 (0.19) -0.048 (0.25)
Ori: SH 0.061 (0.27) 0.410 (0.38)
Ori: AR -0.095 (0.40) 0.068 (0.45)
Ori: AI 0.000 (.) 0.000 (.)
Ori: SG 0.081 (0.14) 0.272 (0.17)
Ori: GR -0.150 (0.21) 0.061 (0.30)
Ori: AG -0.166 (0.15) -0.304 (0.19)
Ori: TG -0.136 (0.20) 0.031 (0.21)
Ori: TI 0.307 (0.25) 0.058 (0.26)
Ori: VD 0.395 (0.27) 0.301 (0.21)
Ori: VS -0.102 (0.19) -0.252 (0.25)
Ori: NE 0.508* (0.29) 0.585** (0.27)
Ori: GE 0.413 (0.32) 0.533* (0.30)
Ori: JU 0.729*** (0.24) 0.452 (0.40)
Constant -0.825*** (0.03) 0.531 (1.03) -0.034 (0.04) 0.535 (1.05) -0.552 (0.43) 0.109 (1.07)
Observations 2817 2760 2817 2808 2817 2805
Pseudo R2 0.007 0.155

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the probit regressions (first two columns) and IV regressions (last four columns)
on unemployment. The data is from the five-year post-graduation surveys. Columns Probit1 and IV1 include no controls, whereas IV3 includes region of origin fixed
effects. Columns Probit2 and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first stage include the
same controls as the estimations from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance at the 1%, 5%
and 10% level, respectively.

26
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A10
Regression table for employment position five years post-graduation.
Oprobit1 Oprobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

A. First stage
Reg. enrol. rate VET 2.501*** (0.24) 1.263 (0.79) -3.625* (1.89) -3.153 (2.93)
B. Second stage
VET 0.141*** (0.04) 0.027 (0.07) 0.783*** (0.16) 0.314 (0.25) -0.904*** (0.17) 0.182 (0.27)
Age 0.041* (0.02) 0.035 (0.02) 0.036 (0.02)
Female -0.122** (0.06) -0.093 (0.06) -0.107* (0.06)
Marital status 0.101 (0.07) 0.094 (0.07) 0.091 (0.07)
Child -0.035 (0.08) -0.034 (0.08) -0.028 (0.08)
Dad: Upp.-sec. -0.189* (0.10) -0.198** (0.10) -0.208** (0.10)
Dad: Post-sec. -0.126 (0.10) -0.132 (0.10) -0.132 (0.10)
Mom: Upp.-sec. 0.219** (0.09) 0.214** (0.09) 0.213** (0.09)
Mom: Post-sec. 0.221** (0.10) 0.225** (0.10) 0.221** (0.10)
Res: BE -0.085 (0.14) -0.090 (0.14) -0.085 (0.15)
Res: LU -0.067 (0.16) -0.046 (0.16) -0.006 (0.18)
Res: UR 0.384 (0.43) 0.397 (0.43) 1.075* (0.61)
Res: SZ 0.108 (0.18) 0.114 (0.18) 0.140 (0.21)
Res: OW 0.008 (0.28) -0.013 (0.29) 0.420 (0.35)
Res: NW 0.018 (0.34) 0.037 (0.34) 0.466 (0.44)
Res: GL -0.640** (0.31) -0.636** (0.30) -0.327 (0.40)
Res: ZG 0.058 (0.16) 0.079 (0.16) 0.310 (0.19)
Res: FR -0.155 (0.19) -0.156 (0.19) -0.022 (0.23)
Res: SO -0.215 (0.17) -0.243 (0.17) -0.408* (0.21)
Res: BS -0.098 (0.19) -0.059 (0.20) -0.083 (0.20)
Res: BL 0.108 (0.18) 0.130 (0.18) 0.073 (0.21)
Res: SH -0.065 (0.21) -0.086 (0.21) 0.079 (0.28)
Res: AR 0.426 (0.34) 0.414 (0.33) 0.703* (0.39)
Res: AI 0.013 (0.27) 0.003 (0.27) -1.352 (0.87)
Res: SG 0.061 (0.12) 0.041 (0.12) 0.178 (0.14)
Res: GR 0.028 (0.25) 0.063 (0.26) 0.035 (0.28)
Res: AG -0.035 (0.11) -0.039 (0.11) 0.011 (0.13)
Res: TG -0.236 (0.17) -0.278 (0.17) -0.252 (0.20)
Res: TI -0.327 (0.27) -0.319 (0.26) -0.325 (0.31)
Res: VD 0.017 (0.18) 0.026 (0.18) 0.112 (0.21)
Res: VS 0.086 (0.20) 0.093 (0.20) 0.206 (0.25)
Res: NE 0.164 (0.22) 0.176 (0.23) 0.400 (0.29)
Res: GE -0.242 (0.23) -0.215 (0.23) 0.057 (0.31)
Res: JU 0.184 (0.27) 0.151 (0.27) 0.039 (0.38)
UAS 0.239 (0.24) 0.036 (0.29) 0.081 (0.31)
Master 0.235** (0.10) 0.250** (0.10) 0.250** (0.10)
UNI 2 -0.181 (0.18) -0.185 (0.18) -0.180 (0.18)
UNI 3 -0.479** (0.19) -0.497*** (0.19) -0.466** (0.20)
UNI 4 0.199 (0.20) 0.180 (0.20) 0.279 (0.21)
UNI 5 -0.166 (0.19) -0.171 (0.19) -0.126 (0.20)
UNI 6 -0.377 (0.48) -0.402 (0.48) -0.363 (0.48)
UNI 7 -0.499** (0.23) -0.519** (0.23) -0.448* (0.24)
UNI 8 0.267 (0.21) 0.285 (0.21) 0.295 (0.22)
UNI 9 -0.254 (0.17) -0.254 (0.17) -0.241 (0.18)
UNI 10 -0.190 (0.26) -0.166 (0.26) -0.136 (0.27)
UNI 11 -0.080 (0.19) -0.039 (0.19) -0.009 (0.21)
UNI 12 -0.206 (0.16) -0.161 (0.16) -0.179 (0.17)
UAS 1 -0.347* (0.20) -0.282 (0.20) -0.282 (0.22)
UAS 2 -0.338* (0.20) -0.238 (0.21) -0.193 (0.23)
UAS 3 -0.272 (0.19) -0.222 (0.19) -0.207 (0.20)
UAS 4 -0.413** (0.19) -0.361* (0.19) -0.312 (0.20)
UAS 5 -0.285 (0.27) -0.234 (0.27) -0.203 (0.28)
UAS 6 -0.490*** (0.19) -0.432** (0.18) -0.401** (0.20)
UAS 7 -0.261 (0.18) -0.200 (0.18) -0.186 (0.19)
UAS 8 0.000 (.)
Arts 0.177 (0.23) 0.209 (0.23) 0.159 (0.24)
Edu. Sciences 0.234 (0.16) 0.213 (0.16) 0.175 (0.17)
Economics 0.642*** (0.16) 0.595*** (0.16) 0.572*** (0.17)
Nat. Sciences 0.145 (0.18) 0.102 (0.18) 0.082 (0.18)
Medi. Sciences 0.428* (0.25) 0.394 (0.25) 0.391 (0.25)
Health 0.308 (0.20) 0.306 (0.20) 0.288 (0.20)
Engineering 0.546*** (0.18) 0.466** (0.19) 0.487** (0.19)
Agri. Sciences 0.297* (0.17) 0.208 (0.18) 0.220 (0.19)
Others 0.150 (0.23) 0.103 (0.23) 0.080 (0.23)
Grade 0.180** (0.07) 0.178** (0.07) 0.184** (0.07)
Cohort=2017 -0.165*** (0.05) -0.140** (0.05) -0.149*** (0.06)
Exchange -0.114* (0.06) -0.118* (0.06) -0.118* (0.06)
Pre-study job 0.165** (0.07) 0.171** (0.07) 0.178** (0.07)
Student internship 0.016 (0.06) 0.043 (0.06) 0.043 (0.06)
Student job 0.164** (0.07) 0.156** (0.07) 0.156** (0.07)
(continued on next page)

27
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A10 (continued )

Oprobit1 Oprobit2 IV1 IV2 IV3 IV4

Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE

Work-based high. edu. 0.223** (0.09) 0.194** (0.09) 0.223** (0.09)

Cont. studies -0.179** (0.08) -0.183** (0.08) -0.180** (0.08)
Add. studies 0.090 (0.08) 0.084 (0.08) 0.088 (0.08)
Cont. education 0.188*** (0.07) 0.186*** (0.07) 0.198*** (0.07)
Medium -0.292*** (0.07) -0.293*** (0.07) -0.286*** (0.07)
Large -0.472*** (0.06) -0.471*** (0.06) -0.467*** (0.06)
Public empl. -0.107 (0.07) -0.101 (0.07) -0.100 (0.07)
Electricity ... 0.037 (0.22) 0.022 (0.22) -0.062 (0.20)
Water Supply ... -0.784 (0.52) -0.667 (0.53) -0.692 (0.53)
Construction -0.007 (0.22) -0.023 (0.22) 0.039 (0.23)
Wholesale ... 0.019 (0.11) -0.009 (0.11) 0.009 (0.12)
Transportation ... -0.008 (0.13) 0.006 (0.14) 0.010 (0.14)
Accommodation ... -0.295 (0.35) -0.282 (0.34) -0.328 (0.34)
Information ... -0.299*** (0.10) -0.300*** (0.10) -0.294*** (0.10)
Financial ... -0.258** (0.10) -0.261** (0.10) -0.264** (0.10)
Real Estate 0.165 (0.31) 0.172 (0.31) 0.198 (0.31)
Scientific ... -0.218*** (0.08) -0.217*** (0.08) -0.214*** (0.08)
Administrative ... -0.062 (0.20) -0.040 (0.20) -0.020 (0.21)
Public ... -0.163 (0.15) -0.162 (0.15) -0.166 (0.15)
Education -0.716*** (0.16) -0.696*** (0.16) -0.740*** (0.16)
Health ... -0.359*** (0.13) -0.338*** (0.13) -0.363*** (0.13)
Arts ... -0.005 (0.20) 0.015 (0.19) 0.004 (0.19)
Other service ... 0.003 (0.15) -0.004 (0.15) -0.003 (0.15)
Extraterritorial ... -0.695** (0.30) -0.686** (0.29) -0.657** (0.30)
Lausanne 0.047 (0.15) 0.056 (0.15) 0.024 (0.15)
Sion 0.064 (0.26) 0.066 (0.25) 0.060 (0.26)
Fribourg 0.239 (0.22) 0.225 (0.22) 0.202 (0.22)
Neuchatel 0.055 (0.23) 0.020 (0.24) 0.040 (0.24)
Biel 0.175 (0.22) 0.181 (0.22) 0.186 (0.22)
Bern 0.143 (0.20) 0.152 (0.19) 0.135 (0.20)
Basel 0.212 (0.24) 0.217 (0.24) 0.220 (0.24)
Aarau-Olten 0.338 (0.22) 0.356 (0.22) 0.332 (0.22)
Zurich 0.220 (0.20) 0.232 (0.20) 0.219 (0.20)
Winterthur-SH 0.377* (0.23) 0.389* (0.22) 0.372 (0.23)
St. Gallen 0.310 (0.23) 0.333 (0.23) 0.311 (0.23)
Chur 0.390 (0.32) 0.353 (0.32) 0.389 (0.32)
Luzern 0.374 (0.24) 0.378 (0.24) 0.352 (0.24)
Bellinzona 0.839** (0.36) 0.839** (0.35) 0.832** (0.35)
Lugano 0.396 (0.35) 0.386 (0.35) 0.389 (0.36)
Abroad 0.375 (0.31) 0.392 (0.31) 0.390 (0.32)
Ori: BE -0.214** (0.09) -0.036 (0.13)
Ori: LU -0.012 (0.11) -0.034 (0.15)
Ori: UR -0.270 (0.30) -0.818* (0.47)
Ori: SZ -0.148 (0.16) -0.049 (0.21)
Ori: OW -0.443** (0.21) -0.596* (0.30)
Ori: NW -0.161 (0.21) -0.562 (0.37)
Ori: GL -0.288 (0.21) -0.421 (0.36)
Ori: ZG -0.484** (0.19) -0.680*** (0.25)
Ori: FR -0.233* (0.12) -0.199 (0.20)
Ori: SO 0.154 (0.12) 0.257 (0.18)
Ori: BS -0.474** (0.22) -0.122 (0.27)
Ori: BL -0.053 (0.13) 0.016 (0.17)
Ori: SH 0.018 (0.15) -0.207 (0.24)
Ori: AR 0.004 (0.24) -0.398 (0.27)
Ori: AI 0.376 (0.49) 1.312 (0.83)
Ori: SG 0.042 (0.09) -0.197 (0.12)
Ori: GR -0.096 (0.15) -0.015 (0.19)
Ori: AG 0.068 (0.09) -0.127 (0.14)
Ori: TG 0.137 (0.12) 0.001 (0.15)
Ori: TI -0.360*** (0.11) -0.070 (0.20)
Ori: VD -0.360*** (0.10) -0.142 (0.18)
Ori: VS -0.245** (0.12) -0.185 (0.21)
Ori: NE -0.422*** (0.14) -0.385* (0.23)
Ori: GE -0.554*** (0.12) -0.439* (0.25)
Ori: JU 0.045 (0.16) 0.041 (0.31)
Observations 2817 2817 2817 2817 2817 2817
Pseudo R2 0.002 0.072

Note: The table shows the regression coefficients (Coef.) and standard errors (SE) from the ordered probit regressions (first two columns) and IV regressions (last four
columns) on employment position. The data is from the five-year post-graduation surveys. Columns Oprobit1 and IV1 include no controls, whereas IV3 includes region
of origin fixed effects. Columns Oprobit2 and IV2 include all controls, whereas IV4 additionally includes region of origin fixed effects. The estimations from the first
stage include the same controls as the estimations from the second stage. Standard errors are robust and no sampling weights are used. ***,** and * denote significance
at the 1%, 5% and 10% level, respectively.

28
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

A3. Estimations from robustness checks

Table A11
Summary statistics of entire dataset and our sample.
Dataset Sample

Mean SE Obs Mean SE Obs

Dependent variables
Wage 39280.70 40518.72 95060 72745.48 23942.50 13899
Search time 13.37 10.42 95060 4.45 5.20 13899
Internship 0.15 0.36 56889 0.17 0.37 13899
Explanatory variables
VET 0.23 0.42 93913 0.38 0.49 13899
Regional enrolment rate in VET 0.70 0.12 65179 0.72 0.11 13899
Channel
Related VET 0.14 0.35 95050 0.27 0.45 13899
CV: Personal characteristics
Age 27.29 4.56 95060 26.19 1.60 13899
Female 0.53 0.50 95060 0.50 0.50 13899
Marital status 0.10 0.30 89051 0.06 0.24 13899
Child 0.06 0.23 89034 0.02 0.15 13899
Education of father
Dad: Comp. 0.08 0.27 86244 0.08 0.27 13899
Dad: Upp.-sec. 0.36 0.48 86244 0.42 0.49 13899
Dad: Post-sec. 0.57 0.50 86244 0.51 0.50 13899
Education of mother
Mom: Comp. 0.10 0.30 86704 0.09 0.29 13899
Mom: Upp.-sec. 0.53 0.50 86704 0.61 0.49 13899
Mom: Post-sec. 0.37 0.48 86704 0.30 0.46 13899
CV: Education
UAS 0.34 0.47 95060 0.62 0.49 13899
Institute
UNI 1 0.06 0.24 95060 0.03 0.16 13899
UNI 2 0.08 0.27 95060 0.04 0.20 13899
UNI 3 0.06 0.23 95060 0.04 0.18 13899
UNI 4 0.07 0.26 95060 0.03 0.16 13899
UNI 5 0.07 0.25 95060 0.03 0.18 13899
UNI 6 0.01 0.12 95060 0.00 0.06 13899
UNI 7 0.02 0.15 95060 0.02 0.12 13899
UNI 8 0.03 0.18 95060 0.02 0.15 13899
UNI 9 0.11 0.32 95060 0.05 0.22 13899
UNI 10 0.02 0.13 95060 0.01 0.09 13899
UNI 11 0.04 0.20 95060 0.03 0.18 13899
UNI 12 0.08 0.28 95060 0.08 0.26 13899
UAS 1 0.04 0.19 95060 0.07 0.26 13899
UAS 2 0.09 0.29 95060 0.16 0.37 13899
UAS 3 0.04 0.21 95060 0.08 0.27 13899
UAS 4 0.03 0.18 95060 0.06 0.25 13899
UAS 5 0.02 0.13 95060 0.02 0.14 13899
UAS 6 0.03 0.17 95060 0.07 0.26 13899
UAS 7 0.08 0.27 95060 0.15 0.35 13899
UAS 8 0.00 0.06 95060 0.00 0.06 13899
Master 0.35 0.48 95060 0.35 0.48 13899
Subject
Humanities 0.08 0.26 95060 0.03 0.17 13899
Arts 0.05 0.23 95060 0.03 0.18 13899
Edu. Sciences 0.18 0.39 95060 0.15 0.35 13899
Law 0.08 0.27 95060 0.00 0.00 13899
Economics 0.19 0.39 95060 0.29 0.46 13899
Nat. Sciences 0.11 0.31 95060 0.06 0.24 13899
Medi. Sciences 0.07 0.25 95060 0.02 0.14 13899
Health 0.04 0.20 95060 0.09 0.29 13899
Engineering 0.06 0.23 95060 0.09 0.29 13899
Agri. Sciences 0.12 0.32 95060 0.22 0.41 13899
Others 0.02 0.15 95060 0.01 0.11 13899
Grade 5.08 0.41 86612 5.09 0.39 13899
No. of semester 6.53 2.39 94767 6.18 1.94 13899
Cohort
2011 0.20 0.40 95060 0.10 0.31 13899
2013 0.26 0.44 95060 0.23 0.42 13899
2015 0.27 0.44 95060 0.30 0.46 13899
2017 0.27 0.44 95060 0.37 0.48 13899
CV: Experience
Pre-study job 0.83 0.38 95060 0.88 0.33 13899
Student internship 0.43 0.50 95011 0.46 0.50 13899
(continued on next page)

29
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A11 (continued )

Dataset Sample

Mean SE Obs Mean SE Obs

Student job 0.86 0.35 95060 0.87 0.33 13899

Work-based high. edu. 0.10 0.30 95060 0.13 0.33 13899
Exchange 0.16 0.37 94993 0.16 0.37 13899

Note: The table shows the unweighted summary statistics including mean, standard deviation (SD), and number of observations (Obs) for the entire dataset and our
sample separately. The dataset is from the one-year post-graduation surveys. The main difference between the dataset and our sample lies on our sample focusing on
higher education graduates entering the labour market with little work experience required in their higher education curricula. See the methodology section for more
information on our sample formation.

Table A12 Table A12 (continued )

Regression table on sample selection. HWage
HWage
Coef. SE
Coef. SE
Edu. Sciences 0.143*** (0.02)
Wage Economics 0.225*** (0.02)
VET 0.068*** (0.01) Nat. Sciences 0.141*** (0.02)
Age 0.018** (0.01) Medi. Sciences 0.239*** (0.03)
Female -0.052*** (0.01) Health 0.233*** (0.03)
Marital status 0.023 (0.01) Engineering 0.200*** (0.03)
Child -0.076*** (0.02) Agri. Sciences 0.238*** (0.02)
Dad: Upp.-sec. 0.000 (0.01) Others 0.026 (0.03)
Dad: Post-sec. 0.002 (0.01) Grade 0.054*** (0.01)
Mom: Upp.-sec. -0.005 (0.01) No. of semester 0.001 (0.00)
Mom: Post-sec. -0.011 (0.01) Cohort=2013 -0.074*** (0.02)
Res: BE -0.055** (0.02) Cohort=2015 -0.085*** (0.02)
Res: LU -0.019 (0.02) Cohort=2017 -0.099*** (0.02)
Res: UR -0.028 (0.06) Exchange -0.002 (0.01)
Res: SZ -0.011 (0.03) Pre-study job 0.002 (0.01)
Res: OW -0.127* (0.05) Student internship 0.036*** (0.01)
Res: NW -0.142** (0.05) Student job -0.006 (0.01)
Res: GL -0.034 (0.06) Work-based high. edu. 0.067*** (0.01)
Res: ZG 0.018 (0.03) Medium 0.081*** (0.01)
Res: FR -0.045 (0.02) Large 0.122*** (0.01)
Res: SO 0.012 (0.02) Public empl. -0.015 (0.01)
Res: BS -0.068** (0.02) Manufacturing 0.299* (0.13)
Res: BL -0.000 (0.03) Electricity ... 0.388** (0.14)
Res: SH -0.051 (0.03) Water Supply ... 0.397** (0.15)
Res: AR -0.036 (0.05) Construction 0.309* (0.14)
Res: AI 0.036 (0.08) Wholesale ... 0.280* (0.13)
Res: SG 0.009 (0.02) Transportation ... 0.222 (0.13)
Res: GR -0.048 (0.03) Accommodation ... -0.030 (0.14)
Res: AG -0.006 (0.02) Information ... 0.301* (0.13)
Res: TG -0.001 (0.02) Financial ... 0.375** (0.13)
Res: TI -0.100*** (0.03) Real Estate 0.253 (0.14)
Res: VD -0.080*** (0.02) Scientific ... 0.306* (0.13)
Res: VS -0.016 (0.03) Administrative ... 0.222 (0.13)
Res: NE -0.108*** (0.03) Public ... 0.218 (0.13)
Res: GE -0.136*** (0.03) Education 0.141 (0.13)
Res: JU -0.068 (0.04) Health ... 0.193 (0.13)
UAS 0.223*** (0.06) Arts ... 0.070 (0.13)
Master 0.162*** (0.01) Other service ... 0.072 (0.13)
UNI 2 0.007 (0.03) Extraterritorial ... 0.081 (0.15)
UNI 3 0.099*** (0.03) Lausanne -0.092*** (0.02)
UNI 4 0.044 (0.03) Sion -0.096** (0.03)
UNI 5 -0.079** (0.03) Fribourg -0.091** (0.03)
UNI 6 0.027 (0.05) Neuchatel -0.077* (0.03)
UNI 7 -0.069* (0.03) Biel -0.039 (0.03)
UNI 8 0.192*** (0.03) Bern -0.051 (0.03)
UNI 9 0.030 (0.02) Basel -0.099** (0.03)
UNI 10 -0.054 (0.04) Aarau-Olten -0.072* (0.03)
UNI 11 0.033 (0.03) Zurich -0.040 (0.03)
UNI 12 0.074** (0.02) Winterthur-SH -0.024 (0.03)
UAS 1 -0.074 (0.05) St. Gallen -0.048 (0.03)
UAS 2 -0.077 (0.06) Chur -0.031 (0.04)
UAS 3 -0.073 (0.06) Luzern -0.039 (0.03)
UAS 4 -0.070 (0.06) Bellinzona -0.039 (0.04)
UAS 5 -0.119 (0.06) Lugano -0.184*** (0.04)
UAS 6 -0.098 (0.06) Abroad -0.251*** (0.04)
UAS 7 -0.086 (0.05) Constant 9.892*** (0.18)
UAS 8 0.000 (.) sample
Arts -0.335*** (0.03) Continued study -18.510 (.)
(continued on next page)

30
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

Table A12 (continued ) Table A12 (continued )

HWage HWage

Coef. SE Coef. SE

Searched employment 7.586*** (0.37) Student job 0.058 (0.03)

VET 0.134*** (0.03) Work-based high. edu. -0.129*** (0.03)
Age -0.358*** (0.01) Medium 0.025 (0.03)
Female 0.000 (0.02) Large -0.039 (0.02)
Marital status -0.156*** (0.03) Public empl. 0.019 (0.03)
Child -0.045 (0.05) Mining ... -4.805 (.)
Dad: Upp.-sec. -0.002 (0.04) Manufacturing 1.026*** (0.29)
Dad: Post-sec. -0.052 (0.04) Electricity ... 0.895** (0.31)
Mom: Upp.-sec. 0.062 (0.04) Water Supply ... 0.743* (0.38)
Mom: Post-sec. -0.097* (0.04) Construction 0.795** (0.31)
Res: BE 0.121* (0.05) Wholesale ... 1.054*** (0.29)
Res: LU 0.033 (0.06) Transportation ... 0.963** (0.30)
Res: UR 0.085 (0.19) Accommodation ... 0.747* (0.31)
Res: SZ 0.099 (0.08) Information ... 0.958** (0.29)
Res: OW -0.143 (0.16) Financial ... 0.854** (0.29)
Res: NW -0.007 (0.16) Real Estate 0.871** (0.32)
Res: GL -0.095 (0.17) Scientific ... 0.816** (0.29)
Res: ZG 0.018 (0.08) Administrative ... 1.063*** (0.30)
Res: FR 0.179* (0.08) Public ... 0.897** (0.30)
Res: SO 0.119 (0.07) Education 0.726* (0.29)
Res: BS 0.080 (0.07) Health ... 0.555 (0.29)
Res: BL 0.196* (0.08) Arts ... 0.954** (0.30)
Res: SH 0.326** (0.12) Other service ... 0.860** (0.30)
Res: AR 0.075 (0.16) Extraterritorial ... 0.463 (0.34)
Res: AI -0.225 (0.23) Lausanne 0.111 (0.06)
Res: SG 0.128* (0.06) Sion -0.001 (0.10)
Res: GR 0.115 (0.09) Fribourg 0.165 (0.09)
Res: AG 0.136** (0.04) Neuchatel 0.214* (0.10)
Res: TG 0.098 (0.07) Biel 0.055 (0.09)
Res: TI 0.478*** (0.09) Bern 0.240** (0.08)
Res: VD 0.033 (0.07) Basel 0.194* (0.09)
Res: VS 0.229** (0.08) Aarau-Olten 0.237** (0.09)
Res: NE 0.050 (0.10) Zurich 0.244** (0.08)
Res: GE 0.045 (0.09) Winterthur-SH 0.182* (0.09)
Res: JU -0.009 (0.13) St. Gallen 0.227* (0.10)
UAS 0.334* (0.15) Chur 0.239* (0.12)
Master -0.213*** (0.03) Luzern 0.181* (0.09)
UNI 2 -0.024 (0.07) Bellinzona 0.151 (0.14)
UNI 3 0.116 (0.08) Lugano -0.001 (0.13)
UNI 4 -0.015 (0.08) Abroad 0.040 (0.11)
UNI 5 -0.017 (0.08) Constant 0.238 (.)
UNI 6 -0.101 (0.14) /mills
UNI 7 0.209* (0.10) lambda -0.034 (0.04)
UNI 8 -0.113 (0.08) Observations 35,012
UNI 9 0.003 (0.07)
UNI 10 -0.533*** (0.11) Note: The table shows the regression coefficients (Coef.) and standard errors
UNI 11 -0.093 (0.08) (SE) from the two-stage Heckman correction model on wage. The data is from
UNI 12 0.127 (0.07) the one-year post-graduation surveys. The standard control variables are
UAS 1 -0.167 (0.14) included and two instruments (continued study and searched employment) as
UAS 2 -0.133 (0.15) exclusion restrictions. No sampling weights are used.
UAS 3 -0.209 (0.14)
UAS 4 0.026 (0.14)
UAS 5 -0.505** (0.16)
UAS 6 -0.171 (0.14)
UAS 7 -0.179 (0.14)
UAS 8 0.000 (.)
Arts 0.034 (0.07)
Edu. Sciences 0.345*** (0.05)
Law -16.438 (.) Table A13
Economics 0.210*** (0.05) Regression table with nearest neighbour matching.
Nat. Sciences 0.306*** (0.06)
Wage
Medi. Sciences -0.083 (0.07)
Health 0.362*** (0.07) Coef./SE
Engineering 0.520*** (0.06)
ATE
Agri. Sciences 0.400*** (0.06)
r1vs0.VET 0.101***
Others 0.162 (0.08)
(0.02)
Grade -0.038 (0.03)
Observations 6436
No. of semester -0.016*** (0.00)
Cohort=2013 0.693*** (0.03) Note: The table shows the regression coefficients
Cohort=2015 0.863*** (0.03) (Coef.) and standard errors (SE) from the nearest
Cohort=2017 1.264*** (0.03) neighbour matching model for wage. The data is
Exchange -0.034 (0.03)
from the one-year post-graduation surveys. We use
Pre-study job 0.119*** (0.03)
the observable variables age, gender, type of edu­
Student internship -0.079*** (0.02)
cation, level of education, subject, and final grades
for matching. No sampling weights are used.

31
M.E. Oswald-Egg and U. Renold Economics of Education Review 80 (2021) 102065

References Curtis, S., & Shani, N. (2002). The effect of taking paid employment during term-time on
students’ academic studies. Journal of Further and Higher Education, 26(2), 129–138.
Dustmann, C., & Meghir, C. (2005). Wages, experience and seniority. Review of Economic
Agarwal, L., Brunello, G., & Rocco, L. (2019). The pathways to college. (IZA discussion
Studies, 72, 77–108.
paper no. 12691).Retrieved from https://ssrn.com/abstract=3475797
Economist, T. (2014). The internship: Generation i. https://www.economist.com/intern
Altonji, J. G., & Blank, R. M. (1999). Race and gender in the labor market. In
ational/2014/09/12/generation-i.
O. Ashenfelter, & D. Card (Eds.), Handbook of labor economics, volume 3, chapter 48
Egg, M. E. (2016). Macht berufserfahrung den unterschied aus? die bedeutung von
(pp. 3143–3259). North Holland.
berufserfahrung für einen reibungslosen übergang in die berufswelt. KOF Analysen,
Altonji, J. G., & Shakotko, R. A. (1987). Do wages rise with job seniority? Review of
4, 31–43.
Economic Studies, 54(3), 437–459.
Ehrenberg, R. G., & Sherman, D. R. (1987). Employment while in college, academic
Angrist, J. D., Imbens, G. W., & Rubin, D. B. (1996). Identification of causal effects using
achievement, and postcollege outcomes: A summary of results. The Journal of human
instrumental variables. Journal of the American Statistical Association, 91(434),
resources, 22(1), 1–23.
444–455.
Forster, A. G., Bol, T., & van de Werfhorst, H. G. (2016). Vocational education and
Angrist, J. D., & Krueger, A. B. (2001). Instrumental variables and the search for
employment over the life cycle. Sociological science, 3, 473–494.
identification: From supply and demand to natural experiments. Journal of Economic
Frölich, M., & Lechner, M. (2010). Exploiting regional treatment intensity for the
Perspectives, 15(4), 69–85.
evaluation of labor market policies. Journal of the American Statistical Association, 105
Arrow, K. J. (1962). The economic implications of learning by doing. The Review of
(491), 1014–1029.
economic studies, 29(3), 155–173.
Geel, R., & Backes-Gellner, U. (2012). Earning while learning: When and how student
Arum, R., & Shavit, Y. (1995). Secondary vocational education and the transition from
employment is beneficial. Labour (Committee on Canadian Labour History), 26(3),
school to work. Sociology of education, 68(3), 187–204.
313–340.
Autor, D. H. (2001). Why do temporary help firms provide free general skills training?
Granovetter, M. S. (1973). The strength of ’weak’ ties. American Journal of Sociology, 78
The quarterly journal of economics, 116(4), 1409–1448.
(6), 1360–1380.
Baert, S., Neyt, B., Omey, E., & Verhaest, D. (2017). Student work, educational
Gruber, H. (1999). Erfahrung als grundlage kompetenten handelns. Bern: Huber.
achievement, and later employment: A dynamic approach. (IZA discussion paper no.
Gullason, E. T. (1990). The effects of full-time versus part-time work experience on
11127).Retrieved from http://hdl.handle.net/10419/174037
earnings in ”high” and ”low atrophy” jobs. Economics of education review, 9(3),
Baert, S., Rotsaert, O., Verhaest, D., & Omey, E. (2016). Student employment and later
229–239.
labour market success: no evidence for higher employment chances. Kyklos :
Häkkinen, I. (2006). Working while enrolled in a university: Does it pay? Labour
Jahrbuch des Instituts fur Geschichte der Medizin an der Universitat Leipzig, 69(3),
economics, 13(2), 167–189.
401–425.
Hampf, F., & Woessmann, L. (2017). Vocational vs. general education and employment
Bartolj, T., & Polanec, S. (2018). Does work harm academic performance of students?
over the life cycle: New evidence from PIAAC. CESifo economic studies, 63(3),
evidence using propensity score matching. Research in higher education, 59(4),
255–269.
401–429.
Hanushek, E. A., Schwerdt, G., Woessmann, L., & Zhang, L. (2017). General education,
Bartus, T., & Roodman, D. (2014). Estimation of multiprocess survival models with cmp.
vocational education, and labor-market outcomes over the lifecycle. Journal of
The Stata journal, 14(4), 756–777.
Human Resources, 52(1), 48–87.
Becker, G. S. (1962). Investment in human capital: A theoretical analysis. The journal of
Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica :
political economy, 70(5), 9–49.
journal of the Econometric Society, 47(1), 153–161.
Becker, G. S. (1964). Human capital: A theoretical and empirical analysis, with special
Heckman, J. J., Lochner, L. J., & Todd, P. E. (2006). Earnings functions, rates of return
reference to education. New York: National Bureau of Economic Research.
and treatment effects: The mincer equation and beyond. In E. A. Hanushek, &
Billett, S. (2008). Learning throughout working life: A relational interdependence
F. Welch (Eds.), Handbook of the economics of education, volume 1, chapter 7 (1st, pp.
between personal and social agency. British Journal of Educational Studies, 56(1),
307–458). Elsevier B.V.
39–58.
Heckman, J. J., Stixrud, J., & Urzua, S. (2006). The effects of cognitive and noncognitive
Billett, S. (2014). Integrating learning experiences across tertiary education and practice
abilities on labor market outcomes and social behavior. Journal of labor economics, 24
settings: A socio-personal account. Educational Research Review, 12, 1–13.
(3), 411–482.
Bills, D. B. (2003). Credentials, signals, and screens: Explaining the relationship between
Helyer, R., & Lee, D. (2014). The role of work experience in the future employability of
schooling and job assignment. Review of educational research, 73(4), 441–469.
higher education graduates. Higher Education Quarterly, 68(3), 348–372.
Blackwell, A., Bowes, L., Harvey, L., Hesketh, A. J., & Knight, P. T. (2001). Transforming
Hoffman, N., & Schwartz, R. (2015). Gold Standard: The Swiss Vocational Education and
work experience in higher education. British educational research journal, 27(3),
Training System (International Comparative Study of Vocational Education Systems).
269–285.
Washington, DC: National Center on Education and the Economy.
Blau, F. D., & Kahn, L. M. (2013). The feasibility and importance of adding measures of
Jung, J. H., & Magrabi, F. M. (1991). Work experience, specific human capital, and
actual experience to cross-sectional data collection. Journal of labor economics, 31
earnings. The Quarterly review of economics and business, 31(1), 15–27.
(S1), 17–58.
Klein, M., & Weiss, F. (2011). Is forcing them worth the effort? benefits of mandatory
Böckerman, P., Haapanen, M., & Jepsen, C. (2018). More skilled, better paid: Labour-
internships for graduates from diverse family backgrounds at labour market entry.
market returns to postsecondary vocational education. Oxford economic papers, 70
Studies in Higher Education, 36(8), 969–987.
(2), 485–508.
Laal, M., & Salamati, P. (2012). Lifelong learning; why do we need it? Procedia - Social
Böckerman, P., Haapanen, M., & Jepsen, C. (2019). Back to school: Labor-market returns
and Behavioral Sciences, 31(2011), 399–403.
to higher vocational schooling. Labour economics, 61, 101758.
Lehnert, P., & Pfister, C. (2018). Firms’ Changes in r&d personnel after the introduction
Bolli, T., Caves, K., & Oswald-Egg, M. E. (2019). Valuable experience: How internships
of universities of applied sciences inswitzerland. (economics of education working
affect university graduates’ income. (KOF working paper no. 459).Retrieved from
paper no. 141).Retrieved from https://www.zora.uzh.ch/id/eprint/172257/1/
https://www.research-collection.ethz.ch/handle/20.500.11850/355960
0141_lhwpaper.pdf
Bolli, T., & Hof, S. (2018). The impact of work-based education on non-cognitive skills.
Lemieux, T. (2006). The ‘mincer equation’ thirty years after schooling, experience, and
Journal of research in personality, 75, 46–58.
earnings. In S. Grossbard (Ed.), Jacob mincer a pioneer of modern labor economics,
Bolli, T., Oswald-Egg, M. E., & Rageth, L. (2019). (KOF Working paper no. 429). Meet the
chapter 11 (pp. 127–145). US Boston, MA.: Springer.
need - The role of vocational education and training for the youth labour market.
Light, A., & Ureta, M. (1995). Early-career work experience and gender wage
Retrieved from https://www.research-collection.ethz.ch/handle/20.500.11850/
differentials. Journal of labor economics, 13(1), 121–154.
129532
Margaryan, S., Saniter, N., Schumann, M., & Siedler, T. (2019). Do internships pay off?
Bolli, T., Rageth, L., & Renold, U. (2019). The social status of vocational education and
the effects of student internships on earnings. (IZA discussion paper no. 12478).
training in switzerland. (KOF working paper no. 451).Retrieved from https://www.
Retrieved from http://ftp.iza.org/dp12478.pdf
research-collection.ethz.ch/handle/20.500.11850/323964
Mincer, J. (1974). Schooling, experience, and earnings. NBER.
Brooks, R., & Youngson, P. L. (2016). Undergraduate work placements: An analysis of the
Neyt, B., Omey, E., Verhaest, D., & Baert, S. (2019). Does student work really affect
effects on career progression. Studies in Higher Education, 41(9), 1563–1578.
educational outcomes? a review of the literature. Journal of economic surveys, 33(3),
Burda, M. (2001). Einige überlegungen zu den folgen endlicher und überlappender
896–921.
lebzeiten für die humankapitaltheorie. In R. von Weizsäcker (Ed.), Bildung und
Nunley, J. M., Pugh, A., Romero, N., & Seals, R. A. (2016). College major, internship
beschäftigung, chapter 284 (pp. 11–24). Berlin: Schriften des Vereins für Socialpolitik.
experience, and employment opportunities: Estimates from a résumé audit. Labour
Burdett, K., Carrillo-Tudela, C., & Coles, M. G. (2011). Human capital accumulation and
economics, 38, 37–46.
labor market equilibrium. International economic review, 52(3), 657–677.
OECD. (2004). OECD Handbook for internationally comparative education statistics.
Card, D., & Krueger, A. B. (1994). Minimum wages and employment: A case study of the
concepts, standards, definitions and classifications. Paris: OECD Publishing.
fast-food industry in new jersey and pensylvania. The American economic review, 84
OECD. (2017). Education policy outlook country profile: Italy. OECD.
(4), 772–793.
Oggenfuss, C., & Wolter, S. C. (2018). Are they coming back ? the mobility of university
Cedefop. (2012). Permeable education and training systems: reducing barriers and
students in switzerland after graduation. (IZA discussion paper no. 11349).Retrieved
increasing opportunity. Technical report november. European Centre for the
from http://ftp.iza.org/dp11349.pdf
Development of Vocational Training (Cedefop), Thessaloniki.
Pauline, M., & Simon, F. (2013). OECD Reviews of vocational education and training: a skills
Cerulli-Harms, A. (2017). Generation internship: The impact of internships on early job
beyond school review of england. OECD Publishing/OECD.
market performance. (IZA discussion paper no. 11163).Retrieved from http://ftp.iza.
Pfister, C., Rinawi, M., Harhoff, D., & Backes-Gellner, U. (2018). Regional innovation
org/dp11163.pdf
effects of applied research institutions. (economics of education working paper no.
Cranmer, S. (2006). Enhancing graduate employability: Best intentions and mixed
117).Retrieved from: http://repec.business.uzh.ch/RePEc/iso/leadinghous
outcomes. Studies in Higher Education, 31(2), 169–184.
e/0117_lhwpaper.pdf

32
M.E. Oswald-Egg and U. Renold
Quiñones, M. A., Ford, J. K., & Teachout, M. S. (1995). The relationship between work
experience and job performance: A conceptual and meta-analytic review. Personnel
psychology, 48(4), 887–910.
Raelin, J. A. (1997). A model of work-based learning. Organization Science, 8(6),
563–578.
Rageth, L., & Renold, U. (2019). The linkage between the education and employment
systems: Ideal types of vocational education and training programs. Journal of
Education Policy, 1–26.
Rauner, F., & Maclean, R. (2008). Handbook of technical and vocational education and
training research. Springer VS.
Regan, T. L., & Oaxaca, R. L. (2009). Work experience as a source of specification error in
earnings models: Implications for gender wage decompositions. Journal of population
economics, 22(2), 463–499.
Roodman, D. (2011). Fitting fully observed recursive mixed-process models with cmp.
The Stata journal, 11(2), 159–206.
Rosen, S. (1972). Learning and experience in the labor market. The Journal of human
resources, 7(3), 326–342.
Rözer, J. J., & Bol, T. (2019). Labour market effects of general and vocational education
over the life-cycle and across time: Accounting for age, period, and cohort effects.
European sociological review, 35(5), 701–717.
Ruhm, C. J. (1997). Is high school employment consumption or investment? Journal of
labor economics, 15(4), 735–776.
Ryan, P. (2001). The school-to-work transition: A cross-national perspective. Journal of
economic literature, 39(1), 34–92.
Salvisberg, A. (2010). Soft skills auf dem arbeitsmarkt: Bedeutung und wandel (dissertation).
Seismo Verlag. Sozialwissenschaften und Gesellschaftsfragen
San, G. (1986). The early labor force experience of college students and their post-college
success. Economics of education review, 5(1), 65–76.
Schmidlin, S., & Witmer, C. (2007). Generation praktikum - mythos oder realität ? eine
analyse der absolventenbefragungen 1991 bis 2005. Technical report. SERI,
Neuchatel.
SERI. (2020). Vocational and professional education and training in Switzerland: Facts
and figures 2020. Technical report. State Secretariat for Education, Research and
Innovation (SERI), Bern.
Shaw, A. J. (2012). The value of work experience in outcomes for students: An
investigation into the importance of work experience in the lives of female
undergraduates and postgraduate job seekers. Journal of Vocational Education and
Training, 64(2), 155–168.
Silva, P., Lopes, B., Costa, M., Melo, A. I., Dias, G. P., Brito, E., & Seabra, D. (2018). The
million-dollar question: Can internships boost employment? Studies in Higher
Education, 43(1), 2–21.
33
Economics of Education Review 80 (2021) 102065
Solon, G., Haider, S. J., & Wooldridge, J. M. (2015). What are we weighting for? Journal
of Human Resources, 50(2), 301–316.
Spence, M. (1973). Job market signaling. The quarterly journal of economics, 87(3),
355–374.
Steedman, H. (2012). Overview of apprenticeship systems and issues. Technical report.
ILO - International Labour Organization.
Stern, D., Mc Million, M., Hopkins, C., & Stone, J. (1990). Work experience for students
in high school and college. Youth & Society, 21(3), 355–389.
Stiglitz, J. E. (1975). The theory of screening, education, and the distribution of income.
The American economic review, 65(3), 283–300.
Thurow, L. C. (1975). Generating inequality. London and Basingstoke: The Macmillan
Press LTD.
Tomlinson, M. (2008). ’The degree is not enough’: Students’ perceptions of the role of
higher education credentials for graduate work and employability. British journal of
sociology of education, 29(1), 49–61.
Tuor, S. N., & Backes-Gellner, U. (2010). Risk-return trade-offs to different educational
paths: Vocational, academic and mixed. Journal of Economic Studies, 31(5), 495–519.
Van Belle, E., Caers, R., Cuypers, L., De Couck, M., Neyt, B., Van Borm, H., & Baert, S.
(2019). What do student jobs on graduates CVs signal to employers?. (GLO discussion
paper no. 356).Retrieved from http://hdl.handle.net/10419/198035, Essen
Weiss, F., Klein, M., & Grauenhorst, T. (2014). The effects of work experience during
higher education on labour market entry: Learning by doing or an entry ticket?
Work, Employment & Society, 28(5), 788–807.
West, M. R., & Woessmann, L. (2010). ’Every catholic child in a catholic school’:
historical resistance to state schooling, contemporary private competition and
student achievement across countries. The Economic Journal, 120, 229–255.
Williams, N. (1991). Reexamining the wage, tenure and experience relationship. The
review of economics and statistics, 73(3), 512–517.
Winkelmann, R. (1996). Employment prospects and skill acquisition of apprenticeship-
trained workers in germany. Industrial and Labor Relations Review, 49(4), 658–672.
Wolter, S. C., Cattaneo, M. A., Denzler, S., Diem, A., Grossenbacher, S., . Hof, S., &
Oggenfuss, C. (2014). Swiss education report 2014. Technical report. SKBF CSRE
Swiss Coordination Centre for Research in Education, Aarau.
Wolter, S. C., & Ryan, P. (2011). Apprenticeship. In E. A. Hanushek, S. Machin, &
L. Woessmann (Eds.), Handbook of the economics of education, volume 3, book section
11 (pp. 521–576). Amsterdam: Elsevier B.V.
Yamaguchi, S. (2010). The effect of match quality and specific experience on career
decisions and wage growth. Labour economics, 17(2), 407–423.
Zveglich, J. E., Rodgers, Y.v.d. M., & Lavina, E. (2019). Expected work experience: A new
human capital measure. (ADB economics working paper no. 570).Retrieved from htt
ps://www.adb.org/sites/default/files/publication/481396/ewp-570-expected-wor
k-experience-human-capital-measure.pdf